CN115660859A

CN115660859A - Full-link automatic testing method and system based on Internet finance

Info

Publication number: CN115660859A
Application number: CN202211569195.2A
Authority: CN
Inventors: 方佳璐
Original assignee: Hangyin Consumer Finance Co ltd
Current assignee: Hangyin Consumer Finance Co ltd
Priority date: 2022-12-08
Filing date: 2022-12-08
Publication date: 2023-01-31

Abstract

According to the full-link automatic testing method and system based on internet finance, when a transaction link disturbance data cluster is identified and perfected, the influence of multi-dimensional description knowledge of the transaction link disturbance data cluster and multi-dimensional description knowledge of other transaction link data clusters related to the transaction link disturbance data cluster is taken into consideration, the multi-dimensional description knowledge of the transaction link data clusters is covered in the obtained integrated description knowledge based on the integrated description knowledge of the multi-dimensional description knowledge of the transaction link data clusters, different disturbance ranges of the transaction link data clusters are adapted, the transaction link disturbance data clusters are identified and perfected according to the integrated description knowledge, and therefore the disturbance range of the transaction link data clusters is reduced, and the completeness of the identified and perfected transaction link data clusters is better.

Description

Full-link automatic testing method and system based on Internet finance

Technical Field

The application relates to the field of data processing, in particular to a full-link automatic testing method and system based on internet finance.

Background

In the financial field, complicated function division relates to a large number of business systems for business processing, transaction systems related to different links complete corresponding business execution contents, and a background needs to perform ordered monitoring on each business link, so that the overall system stability and the automatic error elimination capability are challenged. Before the system is put into operation, in order to verify the system performance, a full link test is required to optimize the system according to a test result. In the testing process, many problems are often found, for example, disturbed link data included in the link test data is not beneficial to subsequent test analysis, and how to effectively process the disturbed link data is a technical problem to be solved.

Disclosure of Invention

The invention aims to provide a full-link automatic testing method and system based on internet finance so as to improve the technical problem.

The technical scheme of the embodiment of the application is realized as follows:

in a first aspect, an embodiment of the present application provides an internet-finance-based full-link automated testing method, which is applied to a financial full-link automated testing system, and the method includes:

responding to the test instruction, and acquiring a financial transaction full link test report; the financial transaction full link test report comprises a plurality of transaction link data clusters which are collected according to a preset period, wherein the transaction link data clusters comprise transaction link disturbance data clusters;

mining first description knowledge of multiple dimensions in each transaction link data cluster, and integrating the first description knowledge corresponding to the same transaction link data cluster to obtain second description knowledge;

according to a common measurement result between the second description knowledge of the transaction link data clusters and the second description knowledge of the transaction link disturbance data clusters, performing eccentric calculation and integration on the second description knowledge of the transaction link data clusters to obtain first integration description knowledge corresponding to the transaction link disturbance data clusters;

identifying and perfecting the transaction link disturbance data cluster according to the first integration description knowledge;

the integrating the multiple first description knowledge corresponding to the same transaction link data cluster to obtain the second description knowledge respectively includes: converting the dimensions of other first description knowledge except the first description knowledge of the target dimension into the target dimension aiming at a plurality of first description knowledge corresponding to any transaction link data cluster; wherein the target dimension is a preset dimension of the plurality of dimensions; and integrating the plurality of converted first description knowledge to obtain second description knowledge of the transaction link data cluster.

As an embodiment, the integrating the converted multiple first descriptive knowledge to obtain the second descriptive knowledge of the transaction link data cluster includes:

performing knowledge splicing on the plurality of converted first description knowledge to obtain splicing description knowledge;

and carrying out linear transformation operation on the splicing description knowledge to obtain second description knowledge of the transaction link data cluster.

As an embodiment, the mining of the first descriptive knowledge of the plurality of dimensions in each transaction link data cluster includes:

mining description knowledge of each transaction link data cluster in the transaction link data clusters to obtain third description knowledge of the transaction link data clusters;

quantizing the third descriptive knowledge to obtain quantized descriptive knowledge of the first dimension;

performing dimension embedding mapping on the quantitative descriptive knowledge of the first dimension to obtain quantitative descriptive knowledge of a second dimension;

information reduction is carried out on the quantitative descriptive knowledge of the second dimension, and first descriptive knowledge of the second dimension is obtained;

and performing dimension extension mapping on the first descriptive knowledge of the second dimension to obtain extended descriptive knowledge of the first dimension, and performing information reduction on the extended descriptive knowledge of the first dimension and the quantitative descriptive knowledge of the first dimension to obtain the first descriptive knowledge of the first dimension.

As one embodiment, said plurality of dimensions includes M, said M > 2, said mining first descriptive knowledge of the plurality of dimensions in each transactional link data cluster includes:

performing dimension embedding mapping and quantization on the quantization description knowledge of the first dimension to obtain quantization description knowledge of a second dimension until obtaining quantization description knowledge of an Nth dimension, wherein N = M-1;

performing dimension embedding mapping on the quantitative description knowledge of the Nth dimension to obtain the quantitative description knowledge of the Mth dimension;

information reduction is carried out on the quantitative descriptive knowledge of the Mth dimension, and first descriptive knowledge of the Mth dimension is obtained;

and performing dimension expansion mapping on the first descriptive knowledge of the Mth dimension to obtain the expanded descriptive knowledge of the Nth dimension, and performing information reduction on the expanded descriptive knowledge of the Nth dimension and the quantized descriptive knowledge of the Nth dimension to obtain the first descriptive knowledge of the Nth dimension until the first descriptive knowledge of the first dimension is obtained.

As an embodiment, the quantification of the first dimension describes the knowledge as a tensor; the dimension embedding mapping and quantization are carried out on the quantitative description knowledge of the first dimension to obtain the quantitative description knowledge of the second dimension, and the method comprises the following steps:

carrying out dimension embedding mapping on the quantitative description knowledge of the first dimension to obtain embedded description knowledge of the second dimension;

disassembling the embedded description knowledge to obtain a plurality of first vector blocks, wherein the first vector blocks comprise description knowledge on a plurality of coordinates;

for each description knowledge, optimizing and adjusting the description knowledge according to a plurality of description knowledge in a first vector block to which the description knowledge belongs and coordinate description knowledge of the plurality of description knowledge, wherein the coordinate description knowledge represents coordinates of the corresponding description knowledge in the first vector block;

establishing a second tensor block by using the plurality of description knowledge after the same first tensor block optimization adjustment;

splicing the plurality of second tensor blocks according to the coordinates of the plurality of first tensor blocks in the embedded description knowledge to obtain a first splicing tensor;

determining a quantitative descriptive knowledge of said second dimension in dependence on said first stitching tensor.

As an embodiment, for each description knowledge, performing optimization adjustment on the description knowledge according to a plurality of description knowledge in a first quantum block to which the description knowledge belongs and coordinate description knowledge of the plurality of description knowledge includes:

for each first vector block, integrating a plurality of description knowledge in the first vector block with corresponding coordinate description knowledge to obtain a plurality of first integrated description knowledge;

for each description knowledge in the first vector block, according to a common measurement result between a first integrated description knowledge corresponding to the description knowledge and the plurality of first integrated description knowledge, performing eccentric calculation and integration on the plurality of first integrated description knowledge, and determining the description knowledge obtained after the eccentric calculation and integration as the description knowledge after optimization and adjustment of the description knowledge;

said determining quantitative descriptive knowledge of said second dimension in dependence on said first splice tensor comprises: determining the first stitching tensor as quantized descriptive knowledge of the second dimension;

before the parsing the embedded description knowledge to obtain a plurality of first scalar partitions, the method further includes: according to the first coordinate conversion tensor, performing coordinate conversion on the description knowledge in the embedding description knowledge to obtain the embedding description knowledge after optimization and adjustment;

said determining quantized descriptive knowledge of said second dimension from said first stitching tensor comprises:

and performing coordinate conversion on the description knowledge in the first splicing tensor according to a second coordinate conversion tensor, determining the description knowledge tensor obtained after conversion as the quantization description knowledge of the second dimension, wherein the second coordinate conversion tensor is an inverse conversion tensor of the first coordinate conversion tensor.

As an embodiment, the information reduction on the extended descriptive knowledge of the nth dimension and the quantized descriptive knowledge of the nth dimension to obtain the first descriptive knowledge of the nth dimension includes:

integrating the extended description knowledge of the Nth dimension with the quantitative description knowledge of the Nth dimension to obtain integrated description knowledge of the Nth dimension;

and carrying out information reduction operation on the integrated description knowledge of the Nth dimension to obtain the first description knowledge of the Nth dimension.

As an embodiment, the full-link automated testing method based on internet finance is executed according to a full-link automated testing network, and the method further includes a debugging step of the full-link automated testing network, including:

acquiring a financial transaction full-link test report sample, wherein the financial transaction full-link test report sample comprises a plurality of transaction link data cluster samples acquired according to a preset period, the transaction link data cluster samples comprise transaction link disturbance data cluster samples, and transaction link data cluster indication information corresponding to the transaction link disturbance data cluster samples is acquired;

processing a transaction link data cluster sample in the financial transaction full link test report sample according to the full link automatic test network to obtain a reasoning transaction link data cluster corresponding to the transaction link disturbance data cluster sample;

according to a data cluster comparison network, mining description knowledge of the transaction link data cluster indication information and the inference transaction link data cluster respectively to obtain fourth description knowledge of the transaction link data cluster indication information and fifth description knowledge of the inference transaction link data cluster;

acquiring a description knowledge error between the fourth description knowledge and the fifth description knowledge according to a data cluster comparison network, and determining the description knowledge error as a first quality evaluation factor;

and debugging the full-link automatic test network according to the first quality evaluation factor.

As an embodiment, after the identification of the transaction link disturbance data cluster according to the first integrated description knowledge is completed, the method further includes:

acquiring the financial transaction full-link test report and a report template library after the financial transaction is identified, wherein the report template library comprises at least one report template corresponding to the transaction type;

mining the description knowledge of the identified financial transaction full link test report to obtain the description knowledge of the financial transaction full link test report, and mining the description knowledge of the report template to obtain the report template description knowledge of the report template;

classifying the identified financial transaction full link test report based on a preset mode according to the financial transaction full link test report description knowledge and the report template description knowledge to obtain a classification result of the identified financial transaction full link test report relative to each transaction type in each preset mode;

and determining the transaction type of the financial transaction full link test report after the completion of the identification according to the classification result of the financial transaction full link test report after the completion of the identification relative to each transaction type under each preset mode.

In a second aspect, an embodiment of the present application provides a financial full-link automation test system, which includes a processor and a memory, where the memory stores a computer program, and the computer program implements the method described above when executed by the processor.

According to the full-link automatic testing method and system based on internet finance, when a transaction link disturbance data cluster is identified and perfected, the influence of multi-dimensional descriptive knowledge of the transaction link disturbance data cluster and multi-dimensional descriptive knowledge of other transaction link data clusters related to the transaction link disturbance data cluster is taken into consideration, the multi-dimensional descriptive knowledge of the transaction link data clusters is covered in the obtained integrated descriptive knowledge based on the integrated descriptive knowledge of the multi-dimensional descriptive knowledge of the transaction link data clusters and is integrated again according to the correlation degree between the transaction link disturbance data clusters, so that different disturbance ranges of the transaction link data clusters are adapted, the transaction link disturbance data clusters are identified and perfected according to the integrated descriptive knowledge, the disturbance range of the transaction link data clusters is relieved and reduced, and the completeness of the identified and perfected transaction link data clusters is better.

In the description that follows, additional features will be set forth, in part, in the description. These features will be in part apparent to those skilled in the art upon examination of the following and the accompanying drawings, or may be learned by production or use. The features of the present application may be realized and attained by practice or use of various aspects of the methodologies, instrumentalities and combinations particularly pointed out in the detailed examples which follow.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the disclosure.

Fig. 1 is a schematic application scenario diagram of a full-link automated testing method based on internet finance according to an embodiment of the present application.

Fig. 2 is a flowchart of a full link automated testing method based on internet finance according to an embodiment of the present application.

Fig. 3 is a schematic diagram of a functional module architecture of a full-link automatic test apparatus according to an embodiment of the present disclosure.

Fig. 4 is a schematic diagram illustrating a financial full-link automation test system according to an embodiment of the present disclosure.

Detailed Description

In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be described in further detail with reference to the attached drawings, the described embodiments should not be considered as limiting the present application, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

In the following description, references to "some embodiments," "as an implementation/scheme," and "in an implementation" describe a subset of all possible embodiments, but it is understood that "some embodiments," "as an implementation/scheme," and "in an implementation" may be the same subset or a different subset of all possible embodiments, and may be combined with each other without conflict.

In the following description, references to the terms "first \ second \ third", and the like, merely distinguish between similar objects and do not denote a particular order, but rather denote a particular order, and it is to be understood that "first \ second \ third", where permissible, may be interchanged either in a particular order or in a sequential order so that embodiments of the application described herein may be practiced in other than that shown or described herein. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the application.

The full-link automatic testing method based on internet finance, provided by the embodiment of the application, can be executed by electronic equipment such as a financial full-link automatic testing system, wherein the electronic equipment can be various types of terminals such as a notebook computer, a tablet computer and a desktop computer, and can also be implemented as a server. The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a Network service, cloud communication, middleware service, a domain name service, a security service, a Content Delivery Network (CDN), a big data and artificial intelligence platform, and the like.

An exemplary application of the financial full-link automation test system device implemented as a server will be described below, and the technical solution in the embodiment of the present application will be clearly and completely described with reference to the drawings in the embodiment of the present application.

Fig. 1 is a schematic application scenario diagram of a full-link automated testing method based on internet finance according to an embodiment of the present application. The plurality of business systems 100 and the financial full-link automated testing system 300 are communicatively connected through a network 200. The financial full-link automated testing system 300 is used for executing the method provided by the embodiment of the application. Specifically, an embodiment of the present application provides an internet finance-based full-link automated testing method, which is applied to a financial full-link automated testing system 300, as shown in fig. 2, and the method includes:

step S10: and responding to the test instruction, and acquiring a financial transaction full-link test report.

In the embodiment of the application, the financial transaction full link test report comprises a plurality of transaction link data clusters acquired according to a preset period, and the transaction link data clusters comprise transaction link disturbance data clusters. The financial transaction full link test report is a financial transaction full link test report obtained by acquiring data in the process of performing a financial transaction full link test, the financial transaction full link test can be a virtual financial transaction test performed in the test process, each link forms a link topology, data in a transaction link data cluster can include but is not limited to transaction objects, transaction types, transaction time, transaction amounts and the like, and after the virtual transaction is completed, the financial transaction full link test report is obtained according to a test instruction. In the financial transaction full link test report, a plurality of transaction link data clusters are sequenced according to the occurrence sequence of transaction behaviors in the transaction link data clusters, and one transaction link data cluster corresponds to one financial transaction link. The data contained in the transaction link disturbance data cluster has disturbance, such as missing, and the transaction link disturbance data cluster is any one of the transaction link data clusters. 4

Step S20: and mining first description knowledge of multiple dimensions in each transaction link data cluster, and integrating the first description knowledge corresponding to the same transaction link data cluster to obtain second description knowledge.

The plurality of dimensions are used for representing dimension values describing knowledge, such as two-dimensional and three-dimensional, the first describing knowledge can be tensor, such as second-order tensor and third-order tensor, and the describing knowledge can be vector representation of data obtained by mining of an artificial intelligence expert system. Each transaction link data cluster corresponds to first descriptive knowledge with multiple dimensions, the first descriptive knowledge with each dimension represents the corresponding transaction link data cluster, and the first descriptive knowledge with different dimensions contains various types of corresponding characteristic information in the transaction link data cluster.

And integrating a plurality of first description knowledge corresponding to the transaction link data clusters to obtain second description knowledge of the transaction link data clusters, so as to obtain the second description knowledge of each transaction link data cluster, wherein each second description knowledge is determined by integrating the first description knowledge of a plurality of dimensions of the corresponding transaction link data clusters, and the second description knowledge covers the multi-dimensional description knowledge of the corresponding transaction link data clusters, so that the information capacity of the second description knowledge is more sufficient.

Step S30: and according to the result of the commonality measurement between the second description knowledge of the transaction link data clusters and the second description knowledge of the transaction link disturbance data clusters, performing eccentric calculation and integration on the second description knowledge of the transaction link data clusters to obtain first integration description knowledge corresponding to the transaction link disturbance data clusters.

In the embodiment of the application, the result of the commonality measurement between the second descriptive knowledge of any transaction link data cluster and the second descriptive knowledge of the transaction link disturbance data cluster represents the correlation between the transaction link data cluster and the transaction link disturbance data cluster. In the application, a plurality of transaction link data clusters acquired according to a preset period in the same report have a relationship, and according to the correlation between each transaction link data cluster and a transaction link disturbance data cluster, second description knowledge of the transaction link data clusters is eccentrically calculated and integrated (i.e. different weights are given to the corresponding second description knowledge according to the correlation, and after the weights are calculated, the second description knowledge is integrated, such as addition or splicing fusion, to obtain first integration description knowledge corresponding to the transaction link disturbance data cluster.

Step S40: and identifying and perfecting the transaction link disturbance data cluster according to the first integration description knowledge.

In the embodiment of the application, because the first integration description knowledge integrates the characteristic information of a plurality of transaction link data clusters, the transaction link disturbance data cluster is identified and perfected based on the first integration description knowledge, the influence of other transaction link data clusters involved in the transaction link disturbance data cluster is comprehensively measured to identify and perfect the transaction link disturbance data cluster, the identified and perfected transaction link data cluster is obtained, disturbance information is eliminated, such as missing filling and redundancy clearing, and subsequent further analysis of the test process, such as classification and analysis of transaction data, is facilitated.

According to the method provided by the steps S10-S40, when the transaction link disturbance data cluster is identified and perfected, the influence of multi-dimensional descriptive knowledge of the transaction link disturbance data cluster and multi-dimensional descriptive knowledge of other transaction link data clusters related to the transaction link disturbance data cluster is measured at the same time, the descriptive knowledge after the multi-dimensional descriptive knowledge of the transaction link data clusters is integrated is adopted, re-integration is carried out according to the correlation degree between the transaction link disturbance data cluster and the acquired integrated descriptive knowledge, the multi-dimensional descriptive knowledge of the transaction link data clusters is covered in the acquired integrated descriptive knowledge, different disturbance ranges of the transaction link data clusters are adapted, the transaction link disturbance data cluster is identified and perfected according to the integrated descriptive knowledge, the disturbance range of the transaction link data cluster is reduced, and the completeness of the identified transaction link data cluster is better.

The above mentioned further analysis of the testing process, such as the analysis of the transaction data classification, is briefly described, wherein for the analysis of the transaction data classification, the analysis may include:

step S51: and acquiring a financial transaction full link test report and a report template library after the completion of the identification.

The report template library comprises at least one report template corresponding to the transaction type.

Step S52: and mining the description knowledge of the identified financial transaction full link test report to obtain the description knowledge of the identified financial transaction full link test report, and mining the description knowledge of the report template to obtain the report template description knowledge of the report template.

And identifying that the complete financial transaction full-link test report description knowledge and the report template description knowledge are description knowledge obtained by mining the description knowledge based on a preset debugging-completed neural network model, wherein the neural network model can be a neural network architecture such as CNN, RNN or LSTM. The method comprises the steps of mining description knowledge of the identified financial transaction full link test report to obtain description knowledge of the financial transaction full link test report, mining description knowledge of a report template, and debugging a neural network model before obtaining the report template description knowledge of the report template, wherein the method specifically comprises the following steps: acquiring a debugging report library, wherein the debugging report library comprises a to-be-estimated debugging report corresponding to annotation transaction type information and at least one debugging report template corresponding to a transaction type; carrying out description knowledge mining on the to-be-estimated debugging report based on the to-be-trained neural network model to obtain description knowledge of the to-be-estimated debugging report, and carrying out description knowledge mining on the debugging report template based on the to-be-trained neural network model to obtain description knowledge of the debugging report template; reasoning the to-be-estimated debugging report according to the to-be-estimated debugging report description knowledge and the debugging report template description knowledge in a preset mode based on the to-be-trained neural network model to obtain a transaction type set of the to-be-estimated debugging report, wherein the transaction type set comprises target transaction types corresponding to each preset mode; and iterating the neural network model to be trained until convergence is achieved through the target transaction type and the transaction type information corresponding to each preset mode, and obtaining the trained neural network model. The method comprises the following steps of obtaining a to-be-estimated debugging report, wherein the to-be-estimated debugging report is inferred according to a preset mode based on a to-be-estimated debugging report description knowledge and a debugging report template description knowledge and a to-be-estimated debugging report model to obtain a transaction type set of the to-be-estimated debugging report, and the method specifically comprises the following steps: calculating matching scores between the to-be-estimated debugging report description knowledge and the debugging report template description knowledge according to a preset mode based on the neural network model to be trained to obtain a temporary matching score between the to-be-estimated debugging report and the debugging report template in each preset mode; determining a target transaction type of a to-be-estimated debugging report in each preset mode according to the temporary matching score; and generating a transaction type set of the predicted debugging report according to the target transaction type of the predicted debugging report in each preset mode.

The method includes the steps of iterating a neural network model to be trained until convergence according to a target transaction type and transaction type information corresponding to each preset mode to obtain a trained neural network model, and specifically may include: acquiring a first error result between a target transaction type corresponding to each preset mode and the transaction type information; integrating the first error result to obtain a target error result; and iterating the neural network model to be trained according to the target error result until convergence, and obtaining the trained neural network model.

Step S53: classifying the financial transaction full link test reports based on preset modes according to the financial transaction full link test report description knowledge and the report template description knowledge to obtain classification results of the financial transaction full link test reports relative to each transaction type in each preset mode.

Specifically, the matching score between the description knowledge of the financial transaction full link test report and the description knowledge of the report template can be determined based on preset modes, and the target matching score between the financial transaction full link test report and the report template corresponding to each transaction type in each preset mode is obtained; and determining the target matching score as a classification result. The preset mode can be various, and the preset mode is a strategy for classifying the financial transaction full link test report, such as cosine distance, euclidean distance, and the like, to determine the target matching score (similarity).

Step S54: and determining the transaction type of the financial transaction full link test report according to the classification result of the financial transaction full link test report relative to each transaction type in each preset mode.

Specifically, the classification results corresponding to each preset mode are integrated with respect to each transaction type to obtain a target classification result corresponding to the transaction type, for example, an eccentricity coefficient corresponding to each preset mode is obtained, the classification results corresponding to each preset mode are integrated based on the eccentricity coefficient to obtain a target classification result corresponding to the transaction type, and then the transaction type of the financial transaction full-link test report is determined according to the target classification result corresponding to the transaction type.

On the basis of the steps S10-S40, the second description knowledge of each transaction link data cluster can be optimized and adjusted or updated according to the second description knowledge of the transaction link data clusters, the sequence relation among the transaction link data clusters is strengthened, and then the transaction link disturbance data clusters are identified and perfected.

As another possible embodiment, the internet finance-based full-link automated testing method provided in the present application may include:

step S100: and responding to the test instruction, and acquiring a financial transaction full-link test report.

The financial transaction full link test report comprises a plurality of transaction link data clusters which are collected according to a preset period, and the transaction link data clusters comprise transaction link disturbance data clusters. In one embodiment, the financial transaction full link test report includes a plurality of transaction link data clusters in succession, and the transaction link disturbance data cluster is a transaction link data cluster in the transaction link data clusters. The transaction link data clusters included in the financial transaction full link test report are arranged in sequence in the transaction link data clusters according to the transaction link data clusters. Aiming at any transaction link data cluster in the report, a plurality of transaction link data clusters before and a plurality of transaction link data clusters after the transaction link data cluster are all involved and connected with the transaction link data cluster, the financial transaction full link test report acquired by the embodiment of the application comprises the transaction link data clusters before and after the transaction link disturbance data cluster so as to accurately identify and perfect the transaction link data cluster. And after the transaction link disturbance data cluster is determined, determining a plurality of transaction link data clusters which are acquired according to a preset period with the transaction link disturbance data cluster according to the transaction link disturbance data cluster, and obtaining a financial transaction full link test report.

As an implementation manner, determining a transaction link disturbance data cluster in a plurality of transaction link data clusters may be to compare each transaction link data cluster with a corresponding preset data evaluation index, and determine whether a field is perfect or whether a redundant field exists, or certainly, the determination may be performed in a manual manner.

Step S200: first descriptive knowledge of a plurality of dimensions is mined in each transactional link data cluster.

Step S300: and converting other first description knowledge except the first description knowledge of the target dimension into the target dimension aiming at a plurality of first description knowledge corresponding to any transaction link data cluster.

In this embodiment, the target dimension is a preset dimension among the multiple dimensions, for any transaction link data cluster, the multiple first descriptive knowledge corresponding to the transaction link data cluster is different multiple dimensions, and other first descriptive knowledge except the first descriptive knowledge of the target dimension among the multiple first descriptive knowledge is converted into the target dimension, so that the converted multiple first descriptive knowledge is the target dimension, and the converted multiple first descriptive knowledge is the same dimension, which is beneficial for integrating the converted multiple first descriptive knowledge.

Step S400: and integrating the plurality of first description knowledge after the conversion of the transaction link data cluster to obtain second description knowledge of the transaction link data cluster.

Because the converted multiple first descriptive knowledge are the same dimension (all target dimensions), the converted multiple first descriptive knowledge can be directly integrated to obtain the second descriptive knowledge of the transaction link data cluster. And processing the plurality of first description knowledge of each transaction link data cluster based on the steps S100-S400 to obtain second description knowledge of each transaction link data cluster.

As an embodiment, step S400 may specifically include: and performing knowledge splicing on the plurality of converted first description knowledge to obtain splicing description knowledge, and performing linear transformation operation (such as filtering smoothing based on a convolution matrix) on the splicing description knowledge to obtain second description knowledge of the transaction link data cluster. Based on this, through knowledge concatenation and convolution, a plurality of first description knowledge after trading link data cluster conversion are handled, can let the description knowledge of a plurality of dimensions of trading link data cluster integrate more fully, and the accuracy of second description knowledge is promoted. Wherein the second knowledge-describing dimension is a target dimension.

Step S500: and aiming at the target description knowledge in the second description knowledge, carrying out eccentric calculation and integration on the second description knowledge according to the common measurement result between the second description knowledge and the target description knowledge to obtain second integrated description knowledge.

The target description knowledge is any one of the plurality of first description knowledge, and the result of the measure of commonality between any one of the second description knowledge and the target description knowledge represents the correlation between the second description knowledge and the target description knowledge, and simultaneously represents the correlation between the transaction link data cluster corresponding to the second description knowledge and the transaction link data cluster corresponding to the target description knowledge. And performing eccentric calculation and integration on the plurality of second descriptive knowledge according to the result of the commonality measurement between the plurality of second descriptive knowledge and the target descriptive knowledge, wherein the second integrated descriptive knowledge obtained by integration simultaneously covers the second descriptive knowledge of the corresponding transaction link data cluster and the descriptive knowledge of other transaction link data clusters involved in the transaction link data cluster.

Step S600: and integrating the second integrated description knowledge with the target description knowledge, and determining the integrated description knowledge as the description knowledge optimized and adjusted by the target description knowledge.

And the description knowledge after the optimization and adjustment of the target description knowledge is the description knowledge after the optimization and adjustment of the transaction link data cluster corresponding to the target description knowledge. And optimizing and updating each second descriptive knowledge based on the steps S500 and S600 to obtain the optimized and adjusted descriptive knowledge of each second descriptive knowledge, so as to obtain the optimized and adjusted descriptive knowledge of a plurality of transaction link data clusters. And optimizing the second description knowledge of each transaction link data cluster one by one according to the second description knowledge of the transaction link data clusters, thereby strengthening the sequential relation among the description knowledge of the transaction link data clusters and the correlation among the description knowledge after the optimization and adjustment of the transaction link data clusters. In addition, the plurality of second descriptive knowledge are eccentrically calculated and integrated to obtain second integrated descriptive knowledge corresponding to the target descriptive knowledge, and the integrated descriptive knowledge is integrated with the target descriptive knowledge, so that the obtained optimized and adjusted descriptive knowledge is adaptive to the transaction link data cluster corresponding to the target descriptive knowledge, and the optimized and adjusted descriptive knowledge is more accurate.

In one embodiment, after obtaining the second descriptive knowledge of the plurality of trade link data clusters, performing optimization adjustment on each second descriptive knowledge through an attention network (e.g., a multi-head attention mechanism) based on steps S500 and S600 to obtain the optimally adjusted descriptive knowledge of each second descriptive knowledge.

Step S700: according to a common measurement result between the description knowledge after the optimization adjustment of the transaction link data clusters and the description knowledge after the optimization adjustment of the transaction link disturbance data clusters, performing eccentric calculation and integration on the description knowledge after the optimization adjustment of the transaction link data clusters to obtain first integration description knowledge corresponding to the transaction link disturbance data clusters.

In the embodiment of the application, the result of the commonality measurement between the optimally adjusted description knowledge of any transaction link data cluster and the optimally adjusted description knowledge of the transaction link disturbance data cluster represents the correlation between the transaction link data cluster and the transaction link disturbance data cluster.

Step S800: and identifying and perfecting the transaction link disturbance data cluster according to the first integration description knowledge.

Because the description knowledge of multiple dimensions of multiple transaction link data clusters is integrated in the first integrated description knowledge, the information capacity of the first integrated description knowledge is more abundant, the transaction link disturbance data clusters are identified and perfected according to the first integrated description knowledge, the influence generated by other transaction link data clusters can be integrated, and the transaction link data clusters obtained by identification and perfection are more accurate. And performing linear transformation operation on the first integration description knowledge to obtain the identified perfect transaction link data cluster.

As an embodiment, step S800 may specifically include: and integrating the first integration description knowledge with the description knowledge after the optimized adjustment of the transaction link disturbance data cluster to obtain third integration description knowledge, and identifying and perfecting the transaction link disturbance data cluster according to the third integration description knowledge. After the first integration description knowledge corresponding to the transaction link disturbance data cluster is obtained, the first integration description knowledge and the description knowledge after the optimization adjustment of the transaction link disturbance data cluster are integrated again, the obtained third integration description knowledge is made to be matched with the transaction link disturbance data cluster, and the third integration description knowledge which can be accurately obtained is ensured, so that the accuracy of identifying the perfect transaction link data cluster is ensured. After obtaining the plurality of second descriptive knowledge optimized and adjusted descriptive knowledge, acquiring the first integrated descriptive knowledge based on the attention network according to step S700, and integrating the first integrated descriptive knowledge with the optimized and adjusted descriptive knowledge of the transaction link disturbance data cluster to obtain accurate third integrated descriptive knowledge. After the second description knowledge of each transaction link data cluster is obtained, the second description knowledge of each transaction link data cluster is adjusted and optimized according to the second description knowledge of the transaction link data clusters, then the first integrated description knowledge is obtained through the second description knowledge after the optimization and adjustment of the transaction link data clusters, or S500 and S600 are skipped, and the second description knowledge of the transaction link data clusters is eccentrically calculated and integrated directly according to the common measurement result between the second description knowledge of the transaction link data clusters and the second description knowledge of the transaction link disturbance data clusters to obtain the first integrated description knowledge. As an embodiment, if the first integrated descriptive knowledge is obtained with the second descriptive knowledge of the plurality of transaction link data clusters, the identifying and refining the transaction link disturbance data cluster comprises: and integrating the first integration description knowledge with the second description knowledge of the transaction link disturbance data cluster to obtain fourth integration description knowledge, and identifying and perfecting the transaction link disturbance data cluster according to the fourth integration description knowledge.

When the transaction link disturbance data cluster is identified and perfected, the influence of multi-dimensional description knowledge of the transaction link disturbance data cluster and multi-dimensional description knowledge of other transaction link data clusters involved in the transaction link disturbance data cluster is simultaneously measured, the multi-dimensional description knowledge of the transaction link data clusters is integrated again according to the description knowledge after the multi-dimensional description knowledge of the transaction link data clusters is integrated, the obtained integrated description knowledge covers the multi-dimensional description knowledge of the transaction link data clusters, so that different disturbance ranges of the transaction link data clusters are adapted, the transaction link disturbance data cluster is identified and perfected according to the integrated description knowledge, the disturbance range of the transaction link data cluster is relieved, and the identified and perfected transaction link data cluster is more accurate.

In addition, after the second descriptive knowledge of each transaction link data cluster is obtained, the second descriptive knowledge of each transaction link data cluster is adjusted and optimized according to the second descriptive knowledge of the transaction link data clusters, sequence continuity among the descriptive knowledge of the transaction link data clusters is increased, the accuracy of the optimized and adjusted descriptive knowledge of each transaction link data cluster is improved, and the accuracy of the transaction link data clusters obtained through perfect identification is improved. In addition, after the first integration description knowledge corresponding to the transaction link disturbance data cluster is obtained, the first integration description knowledge and the description knowledge after the optimization adjustment of the transaction link disturbance data cluster are integrated again, the obtained third integration description knowledge is made to be matched with the transaction link disturbance data cluster, and the accuracy of the obtained third integration description knowledge is guaranteed, so that the accuracy of the transaction link data cluster obtained through complete recognition is guaranteed.

When the plurality of dimensions include M, where M > 2, mining the first descriptive knowledge of the plurality of dimensions in the trading link data clusters according to the ascending order of the plurality of dimensions, as an embodiment, the process of mining the first descriptive knowledge of the plurality of dimensions in each trading link data cluster includes:

(1) And mining description knowledge of the transaction link data cluster aiming at each transaction link data cluster in the plurality of transaction link data clusters to obtain third description knowledge of the transaction link data cluster. Wherein the third descriptive knowledge is indicative of a vectorial representation, such as a tensor (e.g., a second order tensor, a matrix), of the transactional link data cluster.

As an implementation manner, in a large-scale cluster transaction link test scenario, if the data volume contained in each transaction link data cluster is large, the step (1) may specifically include: the method comprises the steps of disassembling a transaction link data cluster to obtain a plurality of data clusters, and mining description knowledge of each data cluster to obtain the description knowledge of each data cluster; and carrying out knowledge splicing on the description knowledge of the data clusters according to the coordinates of the data clusters in the transaction link data cluster to obtain third description knowledge of the transaction link data cluster. And the third description knowledge of the transaction link data cluster is acquired by adopting disassembly and reintegration, so that the description knowledge mining on the large-scale transaction link data cluster is avoided, and the calculation overhead is relieved. As an embodiment, the description knowledge of each data cluster is a tensor (such as a second-order tensor), the description knowledge of the data clusters is spliced into a tensor according to the coordinates of the data clusters in the transaction link data cluster, and the tensor obtained by the knowledge splicing is determined as the third description knowledge of the transaction link data cluster.

(2) And quantizing (e.g., encoding) the third descriptive knowledge to obtain the quantized descriptive knowledge of the first dimension.

In the step (2), the quantitative description knowledge of the first dimension may be a second-order tensor, a third-order tensor, or other tensors of different orders. Because the third descriptive knowledge indicates the feature vector of the transaction link data cluster, quantizing the third descriptive knowledge can strengthen the involvement relationship between the descriptive knowledge in the third descriptive knowledge, and strengthen the expression effect of the quantized descriptive knowledge on the transaction link data cluster. The third descriptive knowledge is a first dimension number, quantification is carried out on the third descriptive knowledge, only the expression effect of the descriptive knowledge on the transaction link data cluster is increased, and the dimension of the descriptive knowledge is not converted.

(3) And performing dimension embedding mapping and quantization on the quantitative descriptive knowledge of the first dimension to obtain the quantitative descriptive knowledge of the second dimension until the quantitative descriptive knowledge of the Nth dimension is obtained, wherein N = M-1.

The method comprises the steps of obtaining quantitative descriptive knowledge of a plurality of dimensions one by one according to a descending order of the plurality of dimensions, wherein the first dimension is a preset dimension in the plurality of dimensions, performing dimension embedding mapping according to the quantitative descriptive knowledge of the first dimension to finish the purpose of dimension reduction, thus obtaining descriptive knowledge of the second dimension, simultaneously performing re-quantization, and strengthening the expression effect of the descriptive knowledge of the second dimension on a transaction link data cluster again. The process of obtaining the quantitative descriptive knowledge of the second dimension is repeated, and the quantitative descriptive knowledge of the next dimension is obtained again until the quantitative descriptive knowledge of the Nth dimension is obtained.

(4) And carrying out dimension embedding mapping on the quantitative description knowledge of the Nth dimension to obtain the quantitative description knowledge of the Mth dimension.

And performing dimension embedding mapping on the quantitative descriptive knowledge of the Nth dimension to obtain the descriptive knowledge of the Mth dimension, and determining the obtained descriptive knowledge of the Mth dimension as the quantitative descriptive knowledge of the Mth dimension.

(5) And carrying out information reduction on the quantitative description knowledge of the Mth dimension to obtain the first description knowledge of the Mth dimension.

And performing information restoration on the Mth quantization description knowledge to strengthen the involvement relation between the description knowledge in the description knowledge so as to strengthen the expression effect of the quantization description knowledge on the transaction link data cluster. When information reduction is carried out on the description knowledge, only the involvement of the description knowledge in the description knowledge is strengthened, and the dimensionality of the description knowledge is not converted.

(6) Performing dimension expansion mapping on the first description knowledge of the Mth dimension to obtain the expansion description knowledge of the Nth dimension; and performing information reduction on the expanded description knowledge of the Nth dimension and the quantitative description knowledge of the Nth dimension to obtain first description knowledge of the Nth dimension until the first description knowledge of the first dimension is obtained.

In the above step, the mth dimension of the M dimensions is a minimum dimension of the M dimensions, the first dimension is a preset dimension of the plurality of dimensions, and the dimensions are gradually decreased from the first dimension to the nth dimension. The first description knowledge of the dimensions is acquired one by one according to the increasing sequence of the dimensions. And simultaneously, the first descriptive knowledge of other dimensions except the first descriptive knowledge of the Mth dimension is acquired according to the first descriptive knowledge of the adjacent smaller dimension and the quantitative descriptive knowledge of the current dimension, so that the expression effect of the first descriptive knowledge of the multiple dimensions on the transaction link data cluster is strengthened. And performing dimension expansion mapping on the first description knowledge of the Mth dimension, acquiring the first description knowledge of the Nth dimension according to the Nth expansion description knowledge and the Nth quantitative description knowledge, and acquiring the first description knowledge of the first dimension based on the rule.

The steps (1) to (6) are described by taking any transaction link data cluster as an example, but for a plurality of transaction link data clusters, first description knowledge of M dimensions of each transaction link data cluster can be obtained according to the steps. Aiming at each transaction link data cluster in the transaction link data clusters, starting from the third description knowledge of the transaction link data cluster, obtaining the first description knowledge of multiple dimensions of each transaction link data cluster through multiple quantization and multiple information reduction, strengthening the expression effect of each first description knowledge on the transaction link data cluster, and increasing the accuracy of the first description knowledge of multiple dimensions of each transaction link data cluster. In addition, the first descriptive knowledge of other dimensions except the first descriptive knowledge of the Mth dimension is determined according to the first descriptive knowledge of the adjacent smaller dimension and the quantitative descriptive knowledge of the current dimension, and the expression effect of the first descriptive knowledge of the multiple dimensions on the transaction link data cluster is strengthened.

In addition, after performing dimension embedding mapping on the quantitative descriptive knowledge of the first dimension, the embedding descriptive knowledge is firstly disassembled into a plurality of tensor blocks for optimization and adjustment, and then is integrated to obtain a second quantitative descriptive knowledge, and the obtaining of the second quantitative descriptive knowledge may specifically include:

(I) And carrying out dimension embedding mapping on the quantitative descriptive knowledge of the first dimension to obtain embedded descriptive knowledge of the second dimension.

In the embodiment of the present application, the two dimensions are smaller than the first dimension, and the dimension embedding mapping for the quantitative description knowledge may be implemented according to any feasible dimension embedding mapping manner, such as an embedding descent method, as an implementation manner, where the second dimension is half of the first dimension.

And (II) disassembling the embedded description knowledge of the second dimension to obtain a plurality of first vector blocks.

And the embedded description knowledge of the second dimension is a tensor, and the embedded description knowledge is disassembled to obtain a plurality of first tensor blocks. Each first vector block includes description knowledge on a plurality of coordinates, for example, the first vector block is a second-order tensor or a third-order tensor, and if a plurality of description knowledge covered by the first vector block is scalars, the first vector block is a second-order tensor, and if a plurality of description knowledge included by the first vector block is vectors, the first vector block is a third-order tensor. As an embodiment, the data size (referred to herein as the size) of each first component block is uniform, in other words, the number of descriptive knowledge on the plurality of coordinates included in each first component block is uniform. In one embodiment, when the embedded descriptive knowledge is a third-order tensor, the plurality of first tensor blocks obtained by decomposing the embedded descriptive knowledge are also third-order tensors, and when the embedded descriptive knowledge is a second-order tensor, the plurality of first tensor blocks obtained by decomposing the embedded descriptive knowledge are also second-order tensors.

And (III) for each description knowledge, optimizing and adjusting the description knowledge according to the description knowledge in the first component block to which the description knowledge belongs and the coordinate description knowledge of the description knowledge.

The coordinate description knowledge characterizes the coordinates of the corresponding description knowledge in the first component block, and the coordinate description knowledge may be a first-order tensor (i.e., vector) or a second-order tensor (i.e., matrix). Aiming at each first vector block, the first vector block comprises a plurality of description knowledge, and each description knowledge is optimized and adjusted according to the description knowledge and the coordinate description knowledge of the description knowledge, so that each description knowledge not only integrates the coordinate description knowledge of itself, but also integrates other description knowledge in the same first vector block and the coordinate description knowledge of other description knowledge, and the involvement relationship between the description knowledge at different positions of the same first vector block is strengthened.

As an embodiment, step (III) specifically includes: integrating a plurality of description knowledge in the first vector blocks with corresponding coordinate description knowledge to obtain a plurality of first integrated description knowledge for each first vector block; and aiming at each description knowledge in the first component block, carrying out eccentric calculation and integration on the plurality of first integrated description knowledge according to a common measurement result between the first integrated description knowledge corresponding to the description knowledge and the plurality of first integrated description knowledge, and determining the description knowledge obtained after the eccentric calculation and integration as the description knowledge after the optimization and adjustment of the description knowledge. And for any descriptive knowledge in the first component blocks, the result of the measurement of the commonality between the first integrated descriptive knowledge corresponding to the descriptive knowledge and the plurality of first integrated descriptive knowledge represents the correlation between the first integrated descriptive knowledge corresponding to the plurality of descriptive knowledge in the first component blocks and the first integrated descriptive knowledge corresponding to the descriptive knowledge. According to the common measurement result between the plurality of first integrated description knowledge and the first integrated description knowledge corresponding to the description knowledge, the plurality of first integrated description knowledge are eccentrically calculated and integrated, the description knowledge obtained after eccentric calculation and integration is determined to be description knowledge after optimization and adjustment of the description knowledge, the description knowledge of the rest positions integrated in the description knowledge after optimization and adjustment is integrated according to the correlation degree between the description knowledge and the rest description knowledge, and therefore the accuracy of the description knowledge after optimization and adjustment is improved.

(IV) building a second tensor block by the same first tensor block optimization adjusted plurality of description knowledge.

And aiming at the same first tensor block, obtaining the description knowledge after the optimization adjustment of the plurality of description knowledge in the first tensor block, namely obtaining the description knowledge after the optimization adjustment of the first tensor block, and establishing a second tensor block by the description knowledge after the optimization adjustment according to the coordinates of the description knowledge in the first tensor block.

And (V) splicing the plurality of second tensor blocks according to the coordinates of the plurality of first tensor blocks in the embedded description knowledge of the second dimension to obtain the first splicing tensor.

The plurality of first tensor blocks are obtained by disassembling the embedded description knowledge, the coordinates of different first tensor blocks in the embedded description knowledge are different, after the second tensor blocks corresponding to the first tensor blocks are obtained, the plurality of second tensor blocks are spliced according to the coordinates of the plurality of first tensor blocks in the embedded description knowledge, the coordinates of the plurality of second tensor blocks in the first splicing tensor are consistent with the coordinates of the corresponding first tensor blocks in the embedded description knowledge, and therefore the obtained first splicing tensor is accurate.

As an embodiment, step (V) may specifically include: and carrying out description knowledge mapping transformation on each second tensor block to obtain an optimally adjusted second tensor block, and splicing the optimally adjusted second tensor blocks according to the coordinates of the first tensor blocks in the embedded description knowledge of the second dimension to obtain a first splicing tensor. After obtaining each second tensor block, performing description knowledge mapping transformation (for example, by linear mapping) on each second tensor block, so that each second tensor block is projected to the target description knowledge value domain to obtain a plurality of optimally adjusted second tensor blocks, in other words, the plurality of optimally adjusted second tensor blocks are located in the target description knowledge value domain. According to the embodiment of the application, the embedding description knowledge is firstly disassembled into the plurality of tensor blocks, the plurality of tensor blocks are optimized and adjusted respectively, then the tensor blocks after optimization and adjustment are subjected to knowledge splicing, and the description knowledge (the first splicing tensor) after the embedding description knowledge is optimized and adjusted is obtained. Because the number of description knowledge covered by each tensor block is less than that of description knowledge covered by the embedded description knowledge, the description knowledge in the tensor blocks is respectively optimized and adjusted after the tensor blocks are disassembled, the number of description knowledge in each tensor block during optimization and adjustment can be less, and the calculation cost is reduced.

(VI) determining the first splice tensor as the quantized descriptive knowledge of the second dimension.

Because the embedded descriptive knowledge is of the second dimension, the optimally adjusted first splice descriptive knowledge is also of the second dimension, and the first splice tensor is a quantized descriptive knowledge of the second dimension. While embodiments of the present application may skip step (VI) by determining the first splice tensor as the quantitative descriptive knowledge of the second dimension directly, other embodiments may determine the quantitative descriptive knowledge of the second dimension based on the first splice tensor. After the first splicing tensor is obtained, the description knowledge mapping transformation is carried out on the first splicing tensor to obtain the quantitative description knowledge of the second dimension. And performing description knowledge mapping transformation on the first splicing tensor so as to project the first splicing tensor to a target description knowledge value domain and obtain quantitative description knowledge of a second dimension.

After the first integration description knowledge is obtained, the first integration description knowledge can be adjusted for multiple times according to the process of optimizing the embedded description knowledge of the second dimension in the steps (II) - (V), and the transaction link disturbance data cluster is identified and perfected according to the optimized and adjusted first integration description knowledge. After the first integration description knowledge is obtained, the first integration description knowledge is adjusted for multiple times, and the expression effect of the first integration description knowledge on the transaction link data cluster is increased.

And when the obtained dimension embedding mapping information of the second dimension is quantized, carrying out coordinate transformation on the description knowledge in the embedding description knowledge, then optimizing the transformed embedding description knowledge through disassembly, and then carrying out coordinate reverse transformation on the description knowledge in the embedding description knowledge after optimization and adjustment, so as to increase the involvement relation among the description knowledge in the finally obtained description knowledge.

The specific step of obtaining the second quantitative descriptive knowledge may specifically include:

(a) And carrying out dimension embedding mapping on the quantitative descriptive knowledge of the first dimension to obtain embedded descriptive knowledge of the second dimension.

(b) And according to the first coordinate conversion tensor, performing coordinate conversion on the description knowledge in the embedded description knowledge to obtain the optimized and adjusted embedded description knowledge.

The first coordinate conversion tensor is used to convert the coordinates describing the knowledge that the tensor contains. In one embodiment, the embedded descriptive knowledge in the second dimension is a second-order tensor, and the tensor obtained by multiplying the first coordinate conversion tensor by the embedded descriptive knowledge is the optimally adjusted embedded descriptive knowledge.

(c) And disassembling the embedding description knowledge after the optimization and the adjustment to obtain a plurality of first scalar blocks.

(d) And for each description knowledge, optimizing and adjusting the description knowledge according to the description knowledge and the coordinate description knowledge of the description knowledge in the first component block to which the description knowledge belongs.

(e) And establishing a second tensor block by the plurality of description knowledge after the same first tensor block optimization adjustment.

(f) And splicing the plurality of second tensor blocks to obtain a first splicing tensor according to the coordinates of the plurality of first tensor blocks in the embedding description knowledge after the optimization and the adjustment.

(g) And performing coordinate conversion on the description knowledge in the first splicing tensor according to the second coordinate conversion tensor, and determining the description knowledge tensor obtained after the conversion as the quantitative description knowledge of the second dimension.

And the second coordinate transformation tensor is an inverse transformation tensor of the first coordinate transformation tensor, after the first splicing tensor is obtained, the coordinate transformation is carried out on the description knowledge in the first splicing tensor according to the second coordinate transformation tensor, the coordinate of each description knowledge in the description knowledge tensor obtained after the transformation is consistent with the coordinate of each description knowledge embedded in the description knowledge before the description knowledge is not adjusted, the reduction of the description knowledge coordinate is completed, and the accuracy of the obtained quantitative description knowledge is guaranteed.

As an embodiment, the step (g) specifically comprises: and carrying out coordinate transformation on the description knowledge in the first splicing tensor, and then carrying out description knowledge mapping transformation on the description knowledge tensor obtained after the transformation to obtain the quantitative description knowledge of the second dimension.

In the process, when the embedded description knowledge is quantized, the description knowledge in the embedded description knowledge is subjected to coordinate conversion to obtain the embedded description knowledge after optimization and adjustment, and then the embedded description knowledge after optimization and adjustment is decomposed into a plurality of first vector blocks to be optimized, so that the description knowledge on other coordinates in different local ranges can be integrated into each description knowledge after optimization and adjustment, the correlation among the description knowledge of different coordinates in the embedded description knowledge is strengthened, and the accuracy of the quantitative description knowledge is improved.

In addition, the specific steps of performing information reduction on the extended descriptive knowledge of the nth dimension and the quantitative descriptive knowledge of the nth dimension may include:

(H1) And integrating the expanded description knowledge of the Nth dimension with the quantitative description knowledge of the Nth dimension to obtain the integrated description knowledge of the Nth dimension.

The expanded descriptive knowledge of the nth dimension and the quantized descriptive knowledge of the nth dimension may both exist in any manner, for example, if the expanded descriptive knowledge of the nth dimension and the quantized descriptive knowledge of the nth dimension are both second-order tensors, the expanded descriptive knowledge of the nth dimension and the quantized descriptive knowledge of the nth dimension are added to obtain a tensor that is the integrated descriptive knowledge of the nth dimension.

(H2) And disassembling the integration description knowledge of the Nth dimension to obtain a plurality of third tensor blocks.

(H3) And for each description knowledge, optimizing and adjusting the description knowledge according to a plurality of description knowledge in the third tensor block to which the description knowledge belongs and the coordinate description knowledge of the plurality of description knowledge.

(H4) And establishing a fourth tensor block by the plurality of description knowledge after the optimization adjustment of the same third tensor block.

(H5) And splicing the plurality of fourth tensor blocks to obtain a second splicing tensor according to the coordinates of the plurality of third tensor blocks in the integration description knowledge of the Nth dimension.

(H6) The second splice tensor is determined as the first descriptive knowledge of the nth dimension.

As an implementation manner, after the second stitching tensor is obtained, the second stitching tensor is subjected to description knowledge mapping transformation to obtain quantitative description knowledge of the nth dimension. And carrying out description knowledge mapping transformation on the second splicing tensor so as to project the second splicing tensor to a target description knowledge value domain and obtain the quantitative description knowledge of the Nth dimension.

On the basis of the steps, before the integrated description knowledge of the Nth dimension is disassembled, coordinate conversion is carried out on the description knowledge in the integrated description knowledge, then tensor blocks are disassembled for quantification, and then the tensor formed by splicing the tensor blocks is subjected to coordinate inverse conversion, so that the correlation degree between the description knowledge in the finally obtained description knowledge is improved. The information recovery of the expanded description knowledge of the Nth dimension and the quantitative description knowledge of the Nth dimension comprises the following steps:

(K1) And integrating the extended description knowledge of the Nth dimension with the quantitative description knowledge of the Nth dimension to obtain the integrated description knowledge of the Nth dimension.

(K2) And according to the third coordinate conversion tensor, performing coordinate conversion on the description knowledge in the integrated description knowledge to obtain the optimized and adjusted integrated description knowledge.

(K3) And disassembling the optimized and adjusted integration description knowledge to obtain a plurality of third tensor blocks.

(K4) And for each description knowledge, optimizing and adjusting the description knowledge according to the description knowledge in the third tensor block to which the description knowledge belongs and the coordinate description knowledge of the description knowledge.

(K5) And forming a fourth tensor block by the plurality of description knowledge after the optimal adjustment of the same third tensor block.

(K6) And splicing the plurality of fourth tensor blocks according to the coordinates of the plurality of third tensor blocks in the optimized and adjusted integrated description knowledge to obtain a second splicing tensor.

(K7) And performing coordinate transformation on the descriptive knowledge in the second splicing tensor according to the fourth coordinate transformation tensor, and determining the descriptive knowledge tensor obtained after the transformation as the first descriptive knowledge of the Nth dimension.

And after the second splicing tensor is obtained, performing coordinate conversion on the description knowledge in the second splicing tensor according to the fourth coordinate conversion tensor, enabling the coordinate of each description knowledge in the description knowledge tensor obtained after conversion to be consistent with the coordinate of each expanded description knowledge in the Nth dimension before optimization of each description knowledge, and finishing the restoration of the coordinate of the description knowledge.

As an embodiment, the step (K7) includes: and after coordinate conversion is carried out on the description knowledge in the second splicing tensor, carrying out description knowledge mapping conversion on the description knowledge tensor obtained after conversion to obtain the quantitative description knowledge of the Nth dimension.

The basis for the description of the above example is M > 2, and in another embodiment, the plurality of dimensions is equal to 2, then mining the first descriptive knowledge of the plurality of dimensions in each transactional link data cluster may specifically include:

(X1) mining description knowledge of the transaction link data cluster aiming at each transaction link data cluster in the transaction link data clusters to obtain third description knowledge of the transaction link data cluster.

And (X2) quantizing the third descriptive knowledge to obtain quantized descriptive knowledge of the first dimension.

And (X3) carrying out dimension embedding mapping on the quantitative description knowledge of the first dimension to obtain the quantitative description knowledge of the second dimension.

And (X4) carrying out information reduction on the quantitative descriptive knowledge of the second dimension to obtain the first descriptive knowledge of the second dimension.

(X5) carrying out dimension expansion mapping on the first description knowledge of the second dimension to obtain the expansion description knowledge of the first dimension, and carrying out information reduction on the expansion description knowledge of the first dimension and the quantization description knowledge of the first dimension to obtain the first description knowledge of the first dimension.

The processing of the transaction link data cluster is executed according to a full-link automatic test network, which needs to be debugged before, and the debugging process specifically includes:

(Y1) acquiring a financial transaction full link test report sample.

The financial transaction full link test report sample comprises a plurality of transaction link data cluster samples collected according to a preset period, the transaction link data cluster samples comprise transaction link disturbance data cluster samples, and transaction link data cluster indication information corresponding to the transaction link disturbance data cluster samples is obtained. The financial transaction full link test report sample is a virtual financial transaction test performed in the test process, each link forms a link topology, and the financial transaction full link test report sample comprises a transaction link disturbance data cluster sample. The data perfection of the transaction link data cluster indication information is greater than that of the transaction link disturbance data cluster sample, for example, the transaction link data cluster indication information is a transaction link data cluster obtained by perfecting the transaction link disturbance data cluster sample, and for example, the transaction link data cluster indication information is obtained by perfecting the transaction link disturbance data cluster sample through manpower.

And (Y2) processing the transaction link data cluster samples in the financial transaction full link test report samples according to the full link automatic test network to obtain inference transaction link data clusters corresponding to the transaction link disturbance data cluster samples.

And (Y3) mining the description knowledge of the transaction link data cluster indication information and the inference transaction link data cluster according to the data cluster comparison network to obtain fourth description knowledge of the transaction link data cluster indication information and fifth description knowledge of the inference transaction link data cluster.

The data cluster comparison network is used for comparing description knowledge of data, the fourth description knowledge is used for indicating transaction link data cluster indication information, and the fifth description knowledge is used for indicating reasoning transaction link data clusters.

(Y4) determining a descriptive knowledge error between the fourth descriptive knowledge and the fifth descriptive knowledge from the data cluster comparison network, the descriptive knowledge error being determined as the first quality assessment factor.

The description knowledge error represents the cost between the fourth description knowledge and the fifth description knowledge, can reflect the precision of the full-link automatic test network according to the description knowledge error, and is determined as a quality evaluation factor (a coefficient for evaluating the precision, such as loss) for debugging the full-link automatic test network.

And (Y5) debugging the full-link automatic test network according to the first quality evaluation factor.

And debugging the full-link automatic test network according to the first quality evaluation factor, and increasing the precision of the full-link automatic test network. As an embodiment, the debugging the full-link automation test network may specifically include: acquiring a second quality evaluation factor according to the data difference between the trading link data cluster indication information and the reasoning trading link data cluster; and debugging the full-link automatic test network according to the first quality evaluation factor and the second quality evaluation factor. And the data difference represents the data difference between the trading link data cluster indication information and the reasoning trading link data cluster. The first quality evaluation factor and the second quality evaluation factor can reflect the precision of the full-link automatic test network, and the full-link automatic test network is debugged according to the first quality evaluation factor and the second quality evaluation factor, so that the precision of the full-link automatic test network is improved.

According to the embodiment of the application, the full-link automatic test network is debugged through the combined data cluster comparison network according to the description knowledge error between the reasoning transaction link data cluster output by the full-link automatic test network and the description knowledge of the indication information of the transaction link data cluster, and the precision of the full-link automatic test network is enhanced. In addition, the data difference of the indication information of the transaction link data cluster and the transaction link data cluster is comprehensively inferred, the full-link automatic test network is debugged by the comprehensive data difference and the description knowledge error, the reference elements are more comprehensive, and the precision of the full-link automatic test network is higher.

Based on the foregoing embodiments, an embodiment of the present application provides a full-link automatic testing apparatus, and fig. 3 is a full-link automatic testing apparatus 340 provided in the embodiment of the present application, and as shown in fig. 3, the apparatus 340 includes:

a report acquiring module 341, configured to acquire a financial transaction full link test report in response to the test instruction; the financial transaction full link test report comprises a plurality of transaction link data clusters which are acquired according to a preset period, wherein the transaction link data clusters comprise transaction link disturbance data clusters;

a knowledge acquisition module 342, configured to mine first descriptive knowledge of multiple dimensions in each transaction link data cluster, and integrate multiple first descriptive knowledge corresponding to the same transaction link data cluster to obtain second descriptive knowledge;

the knowledge integration module 343 is configured to perform eccentric calculation and integration on the second descriptive knowledge of the multiple transaction link data clusters according to a result of commonality measurement between the second descriptive knowledge of the multiple transaction link data clusters and the second descriptive knowledge of the transaction link disturbed data cluster, so as to obtain first integrated descriptive knowledge corresponding to the transaction link disturbed data cluster;

and an identification perfecting module 344, configured to perform identification perfection on the transaction link disturbance data cluster according to the first integration description knowledge.

The above description of the apparatus embodiments, similar to the above description of the method embodiments, has similar beneficial effects as the method embodiments. For technical details not disclosed in the embodiments of the apparatus of the present application, reference is made to the description of the embodiments of the method of the present application for understanding.

If the technical scheme of the application relates to personal or private information, a product applying the technical scheme of the application clearly informs personal information processing rules before processing the personal information, and obtains personal autonomous consent. If the technical scheme of the application relates to sensitive personal information, the product applying the technical scheme of the application obtains individual consent before processing the sensitive personal information, meets the requirement of 'express consent' and is collected in the scope of laws and regulations. For example, at a personal information collection device such as a camera, a clear and significant identifier is set to inform that the personal information collection range is entered, the personal information is collected, and if the person voluntarily enters the collection range, the person is considered as agreeing to collect the personal information; or on the device for processing the personal information, under the condition of informing the personal information processing rule by using obvious identification/information, obtaining personal authorization in the modes of pop-up window information or asking the person to upload personal information thereof and the like; the personal information processing rule may include information such as a personal information processor, a personal information processing purpose, a processing method, and a type of personal information to be processed.

It should be noted that, in the embodiment of the present application, if the alarm processing method is implemented in the form of a software functional module and is sold or used as an independent product, the alarm processing method may also be stored in a computer-readable storage medium. Based on such understanding, the technical solutions of the embodiments of the present application or portions thereof that contribute to the related art may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling an electronic device (which may be a personal computer, a server, or a network device) to execute all or part of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read Only Memory (ROM), a magnetic disk, or an optical disk. Thus, embodiments of the present application are not limited to any specific combination of hardware and software.

An embodiment of the present application provides an electronic device, which includes a memory and a processor, where the memory stores a computer program that can run on the processor, and the processor implements the above-mentioned alarm processing method when executing the computer program.

An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the above-mentioned alarm processing method. The computer readable storage medium may be transitory or non-transitory.

Embodiments of the present application provide a computer program product, which includes a non-transitory computer readable storage medium storing a computer program, and when the computer program is read and executed by a computer, the computer program implements some or all of the steps of the above method. The computer program product may be embodied in hardware, software or a combination thereof. In an alternative embodiment, the computer program product is embodied in a computer storage medium, and in another alternative embodiment, the computer program product is embodied in a Software product, such as a Software Development Kit (SDK), or the like.

It should be noted that fig. 4 is a schematic hardware entity diagram of a financial full-link automation test system provided in an embodiment of the present application, and as shown in fig. 4, the hardware entities of the financial full-link automation test system 300 include: a processor 310, a communication interface 320, and a memory 330, wherein: the processor 310 generally controls the overall operation of the financial full link automated test system 300. The communication interface 320 may enable the electronic device to communicate with other terminals or servers via a network. The Memory 330 is configured to store instructions and applications executable by the processor 310, and may also buffer data (e.g., image data, audio data, voice communication data, and video communication data) to be processed or already processed by the processor 310 and modules in the financial full-link automation test system 300, and may be implemented by a FLASH Memory (FLASH) or a Random Access Memory (RAM). Data may be transferred between processor 310, communication interface 320, and memory 330 via bus 340. Here, it should be noted that: the above description of the storage medium and device embodiments is similar to the description of the method embodiments above, with similar advantageous effects as the method embodiments. For technical details not disclosed in the embodiments of the storage medium and the apparatus of the present application, reference is made to the description of the embodiments of the method of the present application for understanding.

It should be appreciated that reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present application. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. It should be understood that, in the various embodiments of the present application, the sequence numbers of the above-mentioned processes do not mean the execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present application. The above-mentioned serial numbers of the embodiments of the present application are merely for description, and do not represent the advantages and disadvantages of the embodiments.

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising a component of' 8230; \8230;" does not exclude the presence of another like element in a process, method, article, or apparatus that comprises the element.

In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described device embodiments are merely illustrative, for example, the division of the unit is only a logical functional division, and there may be other division ways in actual implementation, such as: multiple units or components may be combined, or may be integrated into another system, or some features may be omitted, or not implemented. In addition, the coupling, direct coupling or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between the devices or units may be electrical, mechanical or other forms.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units; can be located in one place or distributed on a plurality of network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, all the functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may be separately regarded as one unit, or two or more units may be integrated into one unit; the integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.

Those of ordinary skill in the art will understand that: all or part of the steps for realizing the method embodiments can be completed by hardware related to program instructions, the program can be stored in a computer readable storage medium, and the program executes the steps comprising the method embodiments when executed; and the aforementioned storage medium includes: various media that can store program codes, such as a removable Memory device, a Read Only Memory (ROM), a magnetic disk, or an optical disk.

Alternatively, the integrated units described above in the present application may be stored in a computer-readable storage medium if they are implemented in the form of software functional modules and sold or used as independent products. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing an electronic device (which may be a personal computer, a server, or a network device) to execute all or part of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a removable storage device, a ROM, a magnetic or optical disk, or other various media that can store program code.

The above description is only an embodiment of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think of the changes or substitutions within the technical scope of the present application, and shall cover the scope of the present application.

Claims

1. A full-link automatic testing method based on Internet finance is characterized by being applied to a financial full-link automatic testing system, and comprising the following steps:

the integrating the multiple first descriptive knowledge corresponding to the same transaction link data cluster to obtain the second descriptive knowledge respectively includes: converting the dimensions of other first description knowledge except the first description knowledge of the target dimension into the target dimension aiming at a plurality of first description knowledge corresponding to any transaction link data cluster; wherein the target dimension is a preset dimension of the plurality of dimensions; and integrating the plurality of converted first description knowledge to obtain second description knowledge of the transaction link data cluster.

2. The method of claim 1, wherein the integrating the converted plurality of first descriptive knowledge to obtain a second descriptive knowledge of the transaction link data cluster comprises:

3. The method of claim 1, wherein mining the first descriptive knowledge of the plurality of dimensions in each transactional link data cluster comprises:

quantizing the third description knowledge to obtain quantized description knowledge of the first dimension;

performing dimension embedding mapping on the quantitative description knowledge of the first dimension to obtain quantitative description knowledge of a second dimension;

information reduction is carried out on the quantitative description knowledge of the second dimension, and first description knowledge of the second dimension is obtained;

and performing dimension extension mapping on the first description knowledge of the second dimension to obtain extension description knowledge of the first dimension, and performing information reduction on the extension description knowledge of the first dimension and the quantitative description knowledge of the first dimension to obtain the first description knowledge of the first dimension.

4. The method of claim 1, wherein the plurality of dimensions includes M, wherein M > 2, and wherein mining the first descriptive knowledge of the plurality of dimensions in each transactional link data cluster comprises:

performing dimension embedding mapping on the quantitative descriptive knowledge of the Nth dimension to obtain the quantitative descriptive knowledge of the Mth dimension;

and performing dimension extension mapping on the first descriptive knowledge of the M-th dimension to obtain extended descriptive knowledge of the N-th dimension, and performing information reduction on the extended descriptive knowledge of the N-th dimension and the quantitative descriptive knowledge of the N-th dimension to obtain the first descriptive knowledge of the N-th dimension until the first descriptive knowledge of the first dimension is obtained.

5. The method of claim 4, wherein the quantized descriptive knowledge of the first dimension is a tensor; the dimension embedding mapping and quantization are carried out on the quantitative description knowledge of the first dimension to obtain the quantitative description knowledge of the second dimension, and the method comprises the following steps:

for each description knowledge, optimizing and adjusting the description knowledge according to a plurality of description knowledge in a first vector block to which the description knowledge belongs and coordinate description knowledge of the plurality of description knowledge, wherein the coordinate description knowledge represents a coordinate of the corresponding description knowledge in the first vector block;

establishing a second tensor block by the plurality of description knowledge after the same first tensor block is optimized and adjusted;

6. The method according to claim 5, wherein the optimally adjusting the descriptive knowledge according to the descriptive knowledge and the coordinate descriptive knowledge of the descriptive knowledge in the first quantum block to which the descriptive knowledge belongs for each descriptive knowledge comprises:

integrating a plurality of description knowledge in each first vector block with corresponding coordinate description knowledge to obtain a plurality of first integrated description knowledge;

said determining quantized descriptive knowledge of said second dimension from said first stitching tensor comprises: determining the first stitching tensor as quantized descriptive knowledge of the second dimension;

and performing coordinate transformation on the descriptive knowledge in the first splicing tensor according to a second coordinate transformation tensor, determining the descriptive knowledge tensor obtained after the transformation as the quantitative descriptive knowledge of the second dimension, wherein the second coordinate transformation tensor is an inverse transformation tensor of the first coordinate transformation tensor.

7. The method of claim 4, wherein the information recovery of the extended descriptive knowledge of the nth dimension and the quantized descriptive knowledge of the nth dimension to obtain the first descriptive knowledge of the nth dimension comprises:

8. The method according to any one of claims 1 to 7, wherein the full link automated testing method based on the internet finance is executed according to a full link automated testing network, and the method further comprises a debugging step of the full link automated testing network, and the debugging step comprises the following steps:

acquiring financial transaction full-link test report samples, wherein the financial transaction full-link test report samples comprise a plurality of transaction link data cluster samples acquired according to a preset period, the transaction link data cluster samples comprise transaction link disturbance data cluster samples, and transaction link data cluster indication information corresponding to the transaction link disturbance data cluster samples is acquired;

obtaining a description knowledge error between the fourth description knowledge and the fifth description knowledge according to a data cluster comparison network, and determining the description knowledge error as a first quality evaluation factor;

9. The method of claim 8, wherein after identifying the transaction link disturbance data cluster based on the first consolidated descriptive knowledge, the method further comprises:

and determining the transaction type of the identified and improved financial transaction full link test report according to the classification result of the identified and improved financial transaction full link test report relative to each transaction type in each preset mode.

10. A financial full-link automated testing system comprising a processor and a memory, the memory storing a computer program which, when executed by the processor, carries out the method of any one of claims 1 to 9.