CN118467715B - Method, device, equipment and medium for determining associated equipment - Google Patents
Method, device, equipment and medium for determining associated equipment Download PDFInfo
- Publication number
- CN118467715B CN118467715B CN202410936649.8A CN202410936649A CN118467715B CN 118467715 B CN118467715 B CN 118467715B CN 202410936649 A CN202410936649 A CN 202410936649A CN 118467715 B CN118467715 B CN 118467715B
- Authority
- CN
- China
- Prior art keywords
- text
- target
- equipment
- app
- predicted
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000009434 installation Methods 0.000 claims description 39
- 238000004364 calculation method Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 11
- 238000003062 neural network model Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 abstract description 4
- 238000005516 engineering process Methods 0.000 abstract description 2
- 238000012549 training Methods 0.000 description 8
- 238000004458 analytical method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/335—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/31—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Machine Translation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present application relates to the field of text processing technologies, and in particular, to a method, an apparatus, a device, and a medium for determining an associated device, where the method includes: respectively inputting a target usage record text of a target APP in a first device and a target usage record text of a target APP in a second device into a preset large language model, respectively generating a first usage prediction text and a second usage prediction text, acquiring text similarity of the first usage prediction text and the second usage prediction text, and determining that the first device and the second device are associated devices corresponding to the target APP when the text similarity is greater than a preset text similarity threshold; the method and the device can determine the equipment with the association relation from the equipment which is not communicated with the data and can further analyze the association equipment and recommend the push information which accords with the use condition of the equipment.
Description
Technical Field
The present invention relates to the field of text processing technologies, and in particular, to a method, an apparatus, a device, and a medium for determining an associated device.
Background
Currently, in a huge user equipment system, in order to master the use conditions of equipment and application programs of various users, a Software Development Kit (SDK) is adopted to upload the installation information and the use condition information of each APP in the equipment to a database, wherein different new and old equipment used by the same user are not needed, and the equipment and the use condition information of each user are different, so that the analysis of the subsequent use condition of mass equipment is difficult, and therefore, how to determine the relationship between the equipment according to the subsequent use condition of the equipment is a problem to be solved.
Disclosure of Invention
The invention aims to provide a method, a device, equipment and a medium for determining associated equipment, so as to determine equipment with an associated relation in subsequent use conditions, and facilitate subsequent analysis of the associated equipment and recommend pushing information conforming to the use conditions of the equipment.
According to a first aspect of the present invention, there is provided a method of determining an associated device, comprising the steps of:
And inputting the target usage record text of the target APP in the given first device into a preset large language model, and generating a first usage prediction text.
Inputting target usage record text of a target APP in a given second device into a preset large language model, and generating a second usage prediction text; the first device and the second device are determined according to preset classification rules.
And acquiring the text similarity of the first use case predicted text and the second use case predicted text.
When the text similarity is larger than a preset text similarity threshold, determining that the first device and the second device are associated devices corresponding to the target APP.
Further, determining the first device and the second device according to a preset classification rule, including the following steps:
receiving a target usage record text of a target APP in a given initial device and the number of key APP installations corresponding to the initial device; the key APP installation number refers to the number of all APPs installed in the initial equipment.
And according to the preset character quantity weight and the preset installation quantity weight, carrying out weighted sum calculation on the character quantity of the target usage record text and the key APP installation quantity in the initial equipment to obtain the priority of the initial equipment.
And when the priority of the initial equipment is greater than a preset priority threshold, determining the initial equipment as first equipment, and otherwise, determining the initial equipment as second equipment.
Further, the first usage prediction text and the second usage prediction text are prediction texts which are generated by the same prediction text template and have the same format.
Further, the obtaining the text similarity of the first usage prediction text and the second usage prediction text includes the following steps:
According to a plurality of target positions in the predicted text template, a plurality of first predicted values corresponding to the plurality of target positions are obtained from a first service condition predicted text, and a plurality of second predicted values corresponding to the plurality of target positions are obtained from a second service condition predicted text; the target position refers to a position where a missing value exists in the predicted text template.
Obtaining a target ratio corresponding to each target position; the target ratio corresponding to the target position is the ratio of the minimum value to the maximum value in the first predicted value corresponding to the target position and the second predicted value corresponding to the target position.
And according to the given position weight corresponding to each target position, carrying out weighted sum calculation on a plurality of target ratios and a plurality of position weights, and obtaining the text similarity of the first use case predicted text and the second use case predicted text.
Further, the method comprises the following steps:
Comparing the original usage record text of the target APP in the given first device with a given record text template, and judging whether the original usage record text of the target APP in the first device has data missing or not.
When the original usage record text of the target APP in the first device does not have data missing, the original usage record text of the target APP in the first device is determined to be the target usage record text of the target APP in the first device.
When the original usage record text of the target APP in the first device is in data missing, the original usage record text of the target APP in the first device is input into a trained target neural network model, and the target usage record text of the target APP in the first device is acquired.
According to a second aspect of the present invention, there is provided an apparatus for determining an associated device, the apparatus comprising:
The first generation module is used for inputting target usage record text of a target APP in a given first device into a preset large language model and generating a first usage condition prediction text.
The second generation module is used for inputting target usage record text of a target APP in the given second equipment into a preset large language model and generating second usage condition prediction text; the first device and the second device are determined according to preset classification rules.
And the first acquisition module is used for acquiring the text similarity of the first use case predicted text and the second use case predicted text.
The first determining module is used for determining that the first device and the second device are associated devices corresponding to the target APP when the text similarity is larger than a preset text similarity threshold.
Further, the system also comprises a classification module, wherein the classification module comprises:
the receiving module is used for receiving target usage record text of target APP in given initial equipment and the corresponding key APP installation quantity of the initial equipment; the key APP installation number refers to the number of all APPs installed in the initial equipment.
The first calculation module is used for carrying out weighted sum calculation on the character number of the target usage record text and the key APP installation number in the initial equipment according to the preset character number weight and the preset installation number weight, so as to obtain the priority of the initial equipment.
And the second determining module is used for determining the initial equipment as the first equipment when the priority of the initial equipment is greater than a preset priority threshold value, and otherwise, determining the initial equipment as the second equipment.
Further, the first usage prediction text and the second usage prediction text are prediction texts which are generated by the same prediction text template and have the same format.
Further, the first acquisition module includes:
The second obtaining module is used for obtaining a plurality of first predicted values corresponding to the plurality of target positions from the first use condition predicted text according to the plurality of target positions in the predicted text template, and obtaining a plurality of second predicted values corresponding to the plurality of target positions from the second use condition predicted text; the target position refers to a position where a missing value exists in the predicted text template.
The third acquisition module is used for acquiring a target ratio corresponding to each target position; the target ratio corresponding to the target position is the ratio of the minimum value to the maximum value in the first predicted value corresponding to the target position and the second predicted value corresponding to the target position.
And the second calculation module is used for calculating weighted sums of a plurality of target ratios and a plurality of position weights according to the position weights corresponding to each given target position, and obtaining the text similarity of the first use case predicted text and the second use case predicted text.
Further, the method also includes a preprocessing module, the preprocessing module including:
The judging module is used for comparing the original usage record text of the target APP in the given first device with the given record text template and judging whether the original usage record text of the target APP in the first device has data missing or not.
And the third determining module is used for determining the original usage record text of the target APP in the first device as the target usage record text of the target APP in the first device when the original usage record text of the target APP in the first device does not have data missing.
The fourth acquisition module is used for inputting the original usage record text of the target APP in the first equipment into the trained target neural network model when the original usage record text of the target APP in the first equipment is in data missing, and acquiring the target usage record text of the target APP in the first equipment.
According to a third aspect of the present invention, there is provided an electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the method of determining associated devices as described above when executing the computer program.
According to a fourth aspect of the present invention, there is provided a computer readable storage medium storing a computer program which, when executed by a processor, implements the above-described method of determining an associated device.
Compared with the prior art, the invention has at least the following beneficial effects:
According to the method for determining the associated equipment, firstly, target usage record texts of target APP in a given first equipment and target usage record texts of target APP in a second equipment are respectively input into a preset large language model, a first usage prediction text and a second usage prediction text are respectively correspondingly generated, prediction of the usage of the two equipment is achieved, then the text similarity of the first usage prediction text and the second usage prediction text is calculated, when the text similarity is larger than a preset text similarity threshold value, the associated equipment corresponding to the target APP is determined by the first equipment and the second equipment, and equipment with association relation of the subsequent usage can be determined from equipment which is different in type and is not communicated in data through calculating the text similarity of the usage prediction text, so that subsequent analysis of the associated equipment and recommendation of pushing information conforming to the usage of the equipment can be carried out.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for determining an associated device according to a first embodiment of the present invention;
FIG. 2 is a partial flowchart showing steps for obtaining a target usage record text according to an embodiment of the present invention;
FIG. 3 is a flowchart illustrating steps for determining a first device and a second device according to a first embodiment of the present invention;
FIG. 4 is a flowchart of step S300 according to an embodiment of the present invention;
Fig. 5 is a schematic diagram of a determining apparatus of an association device according to a first embodiment of the present invention;
fig. 6 is a schematic diagram of a preprocessing module according to a second embodiment of the present invention;
fig. 7 is a schematic diagram of a classification module according to a second embodiment of the invention;
Fig. 8 is a schematic diagram of a first obtaining module 300 according to a second embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to fall within the scope of the invention.
Example 1
As shown in fig. 1, a first embodiment provides a method for determining an association device, including the following steps:
S100, inputting target usage record text of a target APP in a given first device into a preset large language model, and generating a first usage prediction text.
Specifically, the preset large language model is a model obtained by training according to a plurality of target usage record text samples, and a person skilled in the art knows a training implementation process of the large language model, which is not described herein.
Specifically, the first usage prediction text refers to a text for predicting the usage of the target APP in the first device in a future target time period; the person skilled in the art sets the future target time period according to the actual requirements, for example, predicts the usage of the target APP in the next year.
Specifically, as shown in fig. 2, the method further includes the following steps:
S101, comparing an original usage record text of a target APP in a given first device with a given record text template, and judging whether the original usage record text of the target APP in the first device has data missing or not; it can be understood that: the original use record text takes a record text template as a template to record data, and screening and recording of required data can be carried out according to data information uploaded by sdk.
Preferably, the original usage record text of the target APP comprises average usage time of the target APP in a plurality of different preset periods; the person skilled in the art can also set the required information according to the actual requirements, for example: average number of uses, maximum use time period, or a combination of several information, etc., wherein the target APP may be a shopping APP, etc.
Specifically, the recording text template refers to a template comprising a plurality of pieces of usage data information of the APP to be recorded; for example, the recorded text templates are: the average use time length of the target APP in half year is-the average use time length in one year is-the average use time length in two years is-the average use time length in one year; the average use duration may be an average duration calculated according to a total use duration of each day, or an average duration calculated according to a total duration of each time, which is set by a person skilled in the art according to actual requirements.
S102, when the original usage record text of the target APP in the first device does not have data missing, determining the original usage record text of the target APP in the first device as the target usage record text of the target APP in the first device.
For a better understanding, the following description is made: when the original usage record text of the target APP in the first device is 'the average usage time length of the target APP in half a year is 1 hour, the average usage time length in one year is 1 hour, and the average usage time length in two years is 0.5 hour', no data is lost.
S103, when the original usage record text of the target APP in the first equipment is missing, inputting the original usage record text of the target APP in the first equipment into a trained target neural network model, and acquiring the target usage record text of the target APP in the first equipment; it can be understood that: the data missing of the original usage record text of the target APP refers to all information required by the original usage record text of the target APP, wherein all information required refers to all information required to be recorded, which is provided in a record text template.
For a better understanding, the following description is made: when the original usage record text of the target APP in the first device is 'the average usage time length of the target APP in half a year is 1 hour, the average usage time length in one year is 0.5 hour', and the average usage time length in two years is not recorded, so that data loss exists.
Specifically, the target neural network model is a model obtained by training according to a plurality of usage record text samples with data loss, and a person skilled in the art knows a training implementation process of the neural network model, for example, a natural language processing model based on deep learning, which is not described herein.
The above description is made on the process of obtaining the target usage record text, comparing the original usage record text of the target APP with the record text template, and when the original usage record text does not record the required complete information, performing the completion processing on the recorded information to predict the more accurate usage information in the future time period.
S200, inputting target usage record text of a target APP in a given second device into a preset large language model, and generating a second usage prediction text; the first device and the second device are determined according to preset classification rules.
Preferably, the target usage record text of the target APP in the second device refers to the original usage record text of the target APP in the second device; the second device is considered to be short in service time, and the device performance is perfect, so that the situation of data missing is not considered.
Specifically, the second usage prediction text refers to text for predicting the usage of the target APP in the second device in the future target period.
Specifically, the first usage prediction text and the second usage prediction text are prediction texts which are generated by adopting the same prediction text template and have consistent formats; it can be understood that: the predicted required information types are consistent and the predicted time periods are consistent; for example, the average duration of use of the target APP in the next year is-the average number of uses is-.
According to the above, according to different equipment types, the text recorded with the complete information is selected and input into the trained preset large language model, and the equipment is classified and respectively predicted, so that the predicted text which is more accurate and accords with the actual use condition of the equipment can be obtained.
Preferably, as shown in fig. 3, the determining the first device and the second device according to a preset classification rule includes the following steps:
S001, receiving target usage record text of target APP in given initial equipment and the corresponding key APP installation number of the initial equipment; the key APP installation quantity refers to the quantity of all APPs installed in the initial equipment; if the same APP is repeatedly installed a plurality of times, the number is recorded as an installation number.
S002, according to the preset character quantity weight and the preset installation quantity weight, carrying out weighted sum calculation on the character quantity of the target usage record text and the key APP installation quantity in the initial equipment to obtain the priority of the initial equipment; the person skilled in the art sets the character number weight and the installation number weight according to the actual demand.
Preferably, the character number weight is greater than the installation number weight.
Specifically, the priority of the initial device meets the following conditions:
P=d 1×W1+D2×W2, where P is the priority of the initial device, D 1 is the number of characters of the target usage record text of the target APP in the initial device, W 1 is a preset number of characters weight, D 2 is the number of key APP installations corresponding to the initial device, and W 2 is a preset number of installations weight.
S003, when the priority of the initial equipment is greater than a preset priority threshold, determining that the initial equipment is first equipment, otherwise, determining that the initial equipment is second equipment; the preset priority threshold is set by a person skilled in the art according to the actual requirement, and will not be described herein.
Above-mentioned, on the basis of the target use record text's of target APP character number in initial equipment, also introduced key APP installation quantity to judge the type of equipment, when the priority is higher, indicate that the target use record text's character number and key APP installation quantity are relatively more, regard as old equipment, namely judge it as first equipment, judge it as second equipment on the contrary, through the combination of character number and APP quantity, improved the accuracy of equipment classification.
S300, obtaining the text similarity of the first use case predicted text and the second use case predicted text.
Preferably, as shown in fig. 4, the step S300 further includes the following steps:
S301, according to a plurality of target positions in the predicted text template, a plurality of first predicted values corresponding to the plurality of target positions are obtained from a first service condition predicted text, and a plurality of second predicted values corresponding to the plurality of target positions are obtained from a second service condition predicted text; the target position is a position with a missing value in the predicted text template; it can be understood that: the target position refers to a position where a missing value needs to be added in a predicted text template when the first use condition predicted text or the second use condition predicted text is generated.
For ease of understanding, the predicted text templates and target locations are illustrated: when the predicted text template is 'average using time length of target APP in the next year is' average using times is 'the same', two target positions, namely 'positions in text' are included, when the first use condition prediction text is 'the average use time length of the target APP in the next year is 1h, and the average use times are 3', the first prediction values are 1h and 3 respectively.
S302, obtaining a target ratio corresponding to each target position; the target ratio corresponding to the target position is the ratio of the minimum value to the maximum value in the first predicted value corresponding to the target position and the second predicted value corresponding to the target position.
S303, according to the given position weight corresponding to each target position, carrying out weighted sum calculation on a plurality of target ratios and a plurality of position weights, and obtaining the text similarity of the first use case predicted text and the second use case predicted text.
Preferably, the text similarity meets the following conditions:
S= Σ n j=1 (min(aj,bj)/max(aj,bj)×Hj), where S is the text similarity, n is the number of target positions in the predicted text template, a j is the first predicted value corresponding to the j-th target position in the predicted text template, b j is the second predicted value corresponding to the j-th target position in the predicted text template, H j is the position weight corresponding to the j-th target position in the predicted text template, where the values H 1 to H j decrease in sequence.
According to the method, the data proximity degree of the two predicted values can be reflected through the ratio of the minimum value to the maximum value in the two predicted values, and the importance degree of the data information corresponding to the target position can be reflected through the position weight of the target position, so that the text similarity obtained according to the target ratio and the position weight is more reasonable, and the judgment accuracy of the associated equipment is further improved.
S400, when the text similarity is larger than a preset text similarity threshold, determining that the first device and the second device are associated devices corresponding to a target APP; the text similarity threshold is set by a person skilled in the art according to actual requirements, and will not be described herein.
According to the method for determining the associated equipment, firstly, target usage record texts of target APP in a given first equipment and target usage record texts of target APP in a second equipment are respectively input into a preset large language model, a first usage prediction text and a second usage prediction text are respectively correspondingly generated, prediction of the usage of the two equipment is achieved, then the text similarity of the first usage prediction text and the second usage prediction text is calculated, when the text similarity is larger than a preset text similarity threshold value, the first equipment and the second equipment are determined to be associated equipment corresponding to the target APP, and equipment with association relation in the subsequent usage can be determined from equipment which is different in type and is not communicated in data through calculation of the text similarity of the usage prediction texts, so that subsequent analysis of the associated equipment and recommendation of pushing information conforming to the usage of the equipment can be carried out.
Example two
As shown in fig. 5, the second embodiment provides a determining apparatus for an association device, where the apparatus includes:
The first generating module 100 is configured to input the target usage record text of the target APP in the given first device into a preset large language model, and generate a first usage prediction text.
Specifically, the preset large language model is a model obtained by training according to a plurality of target usage record text samples, and a person skilled in the art knows a training implementation process of the large language model, which is not described herein.
Specifically, the first usage prediction text refers to a text for predicting the usage of the target APP in the first device in a future target time period; the person skilled in the art sets the future target time period according to the actual requirements, for example, predicts the usage of the target APP in the next year.
Specifically, as shown in fig. 6, the apparatus further includes a preprocessing module, where the preprocessing module includes:
A judging module 101, configured to compare an original usage record text of a target APP in a given first device with a given record text template, and judge whether there is a data loss in the original usage record text of the target APP in the first device; it can be understood that: the original use record text takes a record text template as a template to record data, and screening and recording of required data can be carried out according to data information uploaded by sdk.
Preferably, the original usage record text of the target APP comprises average usage time of the target APP in a plurality of different preset periods; the person skilled in the art can also set the required information according to the actual requirements, for example: average number of uses, maximum use time period, or a combination of several information, etc., wherein the target APP may be a shopping APP, etc.
Specifically, the recording text template refers to a template comprising a plurality of pieces of usage data information of the APP to be recorded; for example, the recorded text templates are: the average use time length of the target APP in half year is-the average use time length in one year is-the average use time length in two years is-the average use time length in one year; the average use duration may be an average duration calculated according to a total use duration of each day, or an average duration calculated according to a total duration of each time, which is set by a person skilled in the art according to actual requirements.
A third determining module 102, configured to determine, when there is no data missing in the original usage record text of the target APP in the first device, the original usage record text of the target APP in the first device as the target usage record text of the target APP in the first device.
For a better understanding, the following description is made: when the original usage record text of the target APP in the first device is 'the average usage time length of the target APP in half a year is 1 hour, the average usage time length in one year is 1 hour, and the average usage time length in two years is 0.5 hour', no data is lost.
A fourth obtaining module 103, configured to input, when there is a data loss in an original usage record text of a target APP in the first device, the original usage record text of the target APP in the first device into a trained target neural network model, and obtain a target usage record text of the target APP in the first device; it can be understood that: the data missing of the original usage record text of the target APP refers to all information required by the original usage record text of the target APP, wherein all information required refers to all information required to be recorded, which is provided in a record text template.
For a better understanding, the following description is made: when the original usage record text of the target APP in the first device is 'the average usage time length of the target APP in half a year is 1 hour, the average usage time length in one year is 0.5 hour', and the average usage time length in two years is not recorded, so that data loss exists.
Specifically, the target neural network model is a model obtained by training according to a plurality of usage record text samples with data loss, and a person skilled in the art knows a training implementation process of the neural network model, for example, a natural language processing model based on deep learning, which is not described herein.
The above description is made on the process of obtaining the target usage record text, comparing the original usage record text of the target APP with the record text template, and when the original usage record text does not record the required complete information, performing the completion processing on the recorded information to predict the more accurate usage information in the future time period.
The second generating module 200 is configured to input the target usage record text of the target APP in the given second device into a preset large language model, and generate a second usage prediction text; the first device and the second device are determined according to preset classification rules.
Preferably, the target usage record text of the target APP in the second device refers to the original usage record text of the target APP in the second device; the second device is considered to be short in service time, and the device performance is perfect, so that the situation of data missing is not considered.
Specifically, the second usage prediction text refers to text for predicting the usage of the target APP in the second device in the future target period.
Specifically, the first usage prediction text and the second usage prediction text are prediction texts which are generated by adopting the same prediction text template and have consistent formats; it can be understood that: the predicted required information types are consistent and the predicted time periods are consistent; for example, the average duration of use of the target APP in the next year is-the average number of uses is-.
According to the above, according to different equipment types, the text recorded with the complete information is selected and input into the trained preset large language model, and the equipment is classified and respectively predicted, so that the predicted text which is more accurate and accords with the actual use condition of the equipment can be obtained.
Preferably, as shown in fig. 7, the apparatus further includes a classification module, where the classification module includes:
A receiving module 001, configured to receive a target usage record text of a target APP in a given initial device and a number of key APP installations corresponding to the initial device; the key APP installation quantity refers to the quantity of all APPs installed in the initial equipment; if the same APP is repeatedly installed a plurality of times, the number is recorded as an installation number.
The first calculating module 002 is configured to perform weighted sum calculation on the number of characters of the target usage record text and the number of key APP installations in the initial device according to the preset number of characters and the preset number of installations, so as to obtain a priority of the initial device; the person skilled in the art sets the character number weight and the installation number weight according to the actual demand.
Preferably, the character number weight is greater than the installation number weight.
Specifically, the priority of the initial device meets the following conditions:
P=d 1×W1+D2×W2, where P is the priority of the initial device, D 1 is the number of characters of the target usage record text of the target APP in the initial device, W 1 is a preset number of characters weight, D 2 is the number of key APP installations corresponding to the initial device, and W 2 is a preset number of installations weight.
A second determining module 003, configured to determine, when the priority of the initial device is greater than a preset priority threshold, that the initial device is a first device, and otherwise determine that the initial device is a second device; the preset priority threshold is set by a person skilled in the art according to the actual requirement, and will not be described herein.
Above-mentioned, on the basis of the target use record text's of target APP character number in initial equipment, also introduced key APP installation quantity to judge the type of equipment, when the priority is higher, indicate that the target use record text's character number and key APP installation quantity are relatively more, regard as old equipment, namely judge it as first equipment, judge it as second equipment on the contrary, through the combination of character number and APP quantity, improved the accuracy of equipment classification.
A first obtaining module 300, configured to obtain a text similarity of the first usage prediction text and the second usage prediction text.
Preferably, as shown in fig. 8, the first obtaining module 300 includes:
The second obtaining module 301 is configured to obtain, according to a plurality of target positions in the predicted text template, a plurality of first predicted values corresponding to the plurality of target positions from the first usage prediction text, and obtain a plurality of second predicted values corresponding to the plurality of target positions from the second usage prediction text; the target position is a position with a missing value in the predicted text template; it can be understood that: the target position refers to a position where a missing value needs to be added in a predicted text template when the first use condition predicted text or the second use condition predicted text is generated.
For ease of understanding, the predicted text templates and target locations are illustrated: when the predicted text template is 'average using time length of target APP in the next year is' average using times is 'the same', two target positions, namely 'positions in text' are included, when the first use condition prediction text is 'the average use time length of the target APP in the next year is 1h, and the average use times are 3', the first prediction values are 1h and 3 respectively.
A third obtaining module 302, configured to obtain a target ratio corresponding to each target position; the target ratio corresponding to the target position is the ratio of the minimum value to the maximum value in the first predicted value corresponding to the target position and the second predicted value corresponding to the target position.
And the second calculating module 303 is configured to calculate a weighted sum of a plurality of target ratios and a plurality of position weights according to the given position weights corresponding to each target position, so as to obtain the text similarity of the first usage situation predicted text and the second usage situation predicted text.
Preferably, the text similarity meets the following conditions:
S= Σ n j=1 (min(aj,bj)/max(aj,bj)×Hj), where S is the text similarity, n is the number of target positions in the predicted text template, a j is the first predicted value corresponding to the j-th target position in the predicted text template, b j is the second predicted value corresponding to the j-th target position in the predicted text template, H j is the position weight corresponding to the j-th target position in the predicted text template, where the values H 1 to H j decrease in sequence.
According to the method, the data proximity degree of the two predicted values can be reflected through the ratio of the minimum value to the maximum value in the two predicted values, and the importance degree of the data information corresponding to the target position can be reflected through the position weight of the target position, so that the text similarity obtained according to the target ratio and the position weight is more reasonable, and the judgment accuracy of the associated equipment is further improved.
A first determining module 400, configured to determine that the first device and the second device are associated devices corresponding to a target APP when the text similarity is greater than a preset text similarity threshold; the text similarity threshold is set by a person skilled in the art according to actual requirements, and will not be described herein.
When the determining device of the associated equipment in this embodiment runs, firstly, the target usage record text of the target APP in the given first equipment and the target usage record text of the target APP in the second equipment are respectively input into a preset large language model, a first usage prediction text and a second usage prediction text are respectively generated correspondingly, prediction of the usage of the two equipment is achieved, then, the text similarity of the first usage prediction text and the second usage prediction text is calculated, when the text similarity is larger than a preset text similarity threshold, the associated equipment corresponding to the target APP is determined for the first equipment and the second equipment, and equipment with an associated relation in the subsequent usage can be determined from equipment which is different in type and is not communicated with data through calculating the text similarity of the usage prediction text, so that subsequent analysis can be carried out on the associated equipment and pushing information conforming to the usage of the equipment is recommended.
Example III
The embodiment provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the following steps when executing the computer program:
And inputting the target usage record text of the target APP in the given first device into a preset large language model, and generating a first usage prediction text.
Inputting target usage record text of a target APP in a given second device into a preset large language model, and generating a second usage prediction text; the first device and the second device are determined according to preset classification rules.
And acquiring the text similarity of the first use case predicted text and the second use case predicted text.
When the text similarity is larger than a preset text similarity threshold, determining that the first device and the second device are associated devices corresponding to the target APP.
Example IV
The present embodiment provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of:
And inputting the target usage record text of the target APP in the given first device into a preset large language model, and generating a first usage prediction text.
Inputting target usage record text of a target APP in a given second device into a preset large language model, and generating a second usage prediction text; the first device and the second device are determined according to preset classification rules.
And acquiring the text similarity of the first use case predicted text and the second use case predicted text.
When the text similarity is larger than a preset text similarity threshold, determining that the first device and the second device are associated devices corresponding to the target APP.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SYNCHLINK) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions.
While certain specific embodiments of the invention have been described in detail by way of example, it will be appreciated by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the invention. Those skilled in the art will also appreciate that many modifications may be made to the embodiments without departing from the scope and spirit of the invention. The scope of the invention is defined by the appended claims.
Claims (10)
1. A method of determining an associated device, comprising the steps of:
inputting target usage record text of a target APP in a given first device into a preset large language model, and generating a first usage prediction text;
Inputting target usage record text of a target APP in a given second device into a preset large language model, and generating a second usage prediction text; the first device and the second device are determined according to preset classification rules;
acquiring the text similarity of the first use case predicted text and the second use case predicted text;
When the text similarity is larger than a preset text similarity threshold, determining that the first device and the second device are associated devices corresponding to a target APP;
wherein the first device and the second device are determined according to a preset classification rule, the method comprises the following steps:
Receiving a target usage record text of a target APP in a given initial device and the number of key APP installations corresponding to the initial device; the key APP installation quantity refers to the quantity of all APPs installed in the initial equipment;
According to the preset character quantity weight and the preset installation quantity weight, carrying out weighted sum calculation on the character quantity of the target usage record text and the key APP installation quantity in the initial equipment to obtain the priority of the initial equipment;
and when the priority of the initial equipment is greater than a preset priority threshold, determining the initial equipment as first equipment, and otherwise, determining the initial equipment as second equipment.
2. The method of determining an associated device of claim 1, wherein the first usage prediction text and the second usage prediction text are in-line prediction text generated using a same prediction text template.
3. The method for determining the association device according to claim 2, wherein the obtaining the text similarity of the first usage prediction text and the second usage prediction text includes the steps of:
according to a plurality of target positions in the predicted text template, a plurality of first predicted values corresponding to the plurality of target positions are obtained from a first service condition predicted text, and a plurality of second predicted values corresponding to the plurality of target positions are obtained from a second service condition predicted text; the target position is a position with a missing value in the predicted text template;
obtaining a target ratio corresponding to each target position; the target ratio corresponding to the target position is the ratio of the minimum value to the maximum value in the first predicted value corresponding to the target position and the second predicted value corresponding to the target position;
And according to the given position weight corresponding to each target position, carrying out weighted sum calculation on a plurality of target ratios and a plurality of position weights, and obtaining the text similarity of the first use case predicted text and the second use case predicted text.
4. The method of determining an associated device according to claim 1, further comprising the steps of:
comparing the original usage record text of the target APP in the given first equipment with a given record text template, and judging whether the original usage record text of the target APP in the first equipment has data missing or not;
when the original usage record text of the target APP in the first device does not have data missing, determining the original usage record text of the target APP in the first device as the target usage record text of the target APP in the first device;
When the original usage record text of the target APP in the first device is in data missing, the original usage record text of the target APP in the first device is input into a trained target neural network model, and the target usage record text of the target APP in the first device is acquired.
5. An apparatus for determining an associated device, the apparatus comprising:
The first generation module is used for inputting target usage record text of a target APP in a given first device into a preset large language model and generating a first usage condition prediction text;
The second generation module is used for inputting target usage record text of a target APP in the given second equipment into a preset large language model and generating second usage condition prediction text; the first device and the second device are determined according to preset classification rules;
the first acquisition module is used for acquiring the text similarity of the first use case predicted text and the second use case predicted text;
The first determining module is used for determining that the first device and the second device are associated devices corresponding to a target APP when the text similarity is larger than a preset text similarity threshold;
the apparatus further includes a classification module, the classification module including:
the receiving module is used for receiving target usage record text of target APP in given initial equipment and the corresponding key APP installation quantity of the initial equipment; the key APP installation quantity refers to the quantity of all APPs installed in the initial equipment;
The first calculation module is used for carrying out weighted sum calculation on the character number of the target usage record text and the key APP installation number in the initial equipment according to the preset character number weight and the preset installation number weight to obtain the priority of the initial equipment;
And the second determining module is used for determining the initial equipment as the first equipment when the priority of the initial equipment is greater than a preset priority threshold value, and otherwise, determining the initial equipment as the second equipment.
6. The apparatus according to claim 5, wherein the first usage prediction text and the second usage prediction text are in a consistent format generated using the same prediction text template.
7. The apparatus according to claim 6, wherein the first acquisition module includes:
The second obtaining module is used for obtaining a plurality of first predicted values corresponding to the plurality of target positions from the first use condition predicted text according to the plurality of target positions in the predicted text template, and obtaining a plurality of second predicted values corresponding to the plurality of target positions from the second use condition predicted text; the target position is a position with a missing value in the predicted text template;
The third acquisition module is used for acquiring a target ratio corresponding to each target position; the target ratio corresponding to the target position is the ratio of the minimum value to the maximum value in the first predicted value corresponding to the target position and the second predicted value corresponding to the target position;
And the second calculation module is used for calculating weighted sums of a plurality of target ratios and a plurality of position weights according to the position weights corresponding to each given target position, and obtaining the text similarity of the first use case predicted text and the second use case predicted text.
8. The apparatus for determining an associated device according to claim 5, further comprising a preprocessing module, the preprocessing module comprising:
the judging module is used for comparing the original usage record text of the target APP in the given first equipment with the given record text template and judging whether the original usage record text of the target APP in the first equipment has data missing or not;
A third determining module, configured to determine, when there is no data missing in an original usage record text of a target APP in the first device, the original usage record text of the target APP in the first device as a target usage record text of the target APP in the first device;
The fourth acquisition module is used for inputting the original usage record text of the target APP in the first equipment into the trained target neural network model when the original usage record text of the target APP in the first equipment is in data missing, and acquiring the target usage record text of the target APP in the first equipment.
9. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the method of determining an associated device as claimed in any one of claims 1 to 4 when the computer program is executed.
10. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the method of determining an associated device according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410936649.8A CN118467715B (en) | 2024-07-12 | 2024-07-12 | Method, device, equipment and medium for determining associated equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410936649.8A CN118467715B (en) | 2024-07-12 | 2024-07-12 | Method, device, equipment and medium for determining associated equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118467715A CN118467715A (en) | 2024-08-09 |
| CN118467715B true CN118467715B (en) | 2024-09-27 |
Family
ID=92165402
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410936649.8A Active CN118467715B (en) | 2024-07-12 | 2024-07-12 | Method, device, equipment and medium for determining associated equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118467715B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021189851A1 (en) * | 2020-09-03 | 2021-09-30 | 平安科技（深圳）有限公司 | Text error correction method, system and device, and readable storage medium |
Family Cites Families (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR102325724B1 (en) * | 2015-02-28 | 2021-11-15 | 삼성전자주식회사 | Synchronization of Text Data among a plurality of Devices |
| US10963819B1 (en) * | 2017-09-27 | 2021-03-30 | Amazon Technologies, Inc. | Goal-oriented dialog systems and methods |
| CN108200114B (en) * | 2017-11-22 | 2021-09-14 | 广州广日电气设备有限公司 | Method and system for associating equipment information and control system thereof |
| CN109961080B (en) * | 2017-12-26 | 2022-09-23 | 腾讯科技（深圳）有限公司 | Terminal identification method and device |
| CN110035116A (en) * | 2019-03-13 | 2019-07-19 | 阿里巴巴集团控股有限公司 | The method and apparatus of user-association |
| CN112488140B (en) * | 2019-09-12 | 2024-08-02 | 北京国双科技有限公司 | Data association method and device |
| CN110866241A (en) * | 2019-10-08 | 2020-03-06 | 北京百度网讯科技有限公司 | Evaluation model generation and equipment association method, device and storage medium |
| CN111428021B (en) * | 2020-06-05 | 2023-05-30 | 平安国际智慧城市科技股份有限公司 | Text processing method, device, computer equipment and medium based on machine learning |
| CN113112308B (en) * | 2021-05-06 | 2024-03-12 | 支付宝(杭州)信息技术有限公司 | Equipment labeling method and system and directional information pushing method and system |
| CN114238573B (en) * | 2021-12-15 | 2023-09-22 | 平安科技（深圳）有限公司 | Text countercheck sample-based information pushing method and device |
| CN114492452A (en) * | 2021-12-24 | 2022-05-13 | 深圳云天励飞技术股份有限公司 | Method, device and equipment for training and appealing switching of pre-training language model |
| CN116915831A (en) * | 2023-07-24 | 2023-10-20 | 百度(中国)有限公司 | Sample data generation, model training and method for determining equipment association relation |
| CN117473993A (en) * | 2023-11-09 | 2024-01-30 | 中国工商银行股份有限公司 | Virtual resource acquisition behavior prediction method, device, equipment and storage medium |
| CN117708439B (en) * | 2024-02-06 | 2024-05-31 | 每日互动股份有限公司 | Target text pushing method, device, medium and equipment |
-
2024
- 2024-07-12 CN CN202410936649.8A patent/CN118467715B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021189851A1 (en) * | 2020-09-03 | 2021-09-30 | 平安科技（深圳）有限公司 | Text error correction method, system and device, and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118467715A (en) | 2024-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106991085B (en) | Entity abbreviation generation method and device | |
| CN109783785B (en) | Method and device for generating experiment detection report and computer equipment | |
| CN111797629B (en) | Method and device for processing medical text data, computer equipment and storage medium | |
| CN111783993A (en) | Intelligent labeling method and device, intelligent platform and storage medium | |
| CN112084334B (en) | Label classification method and device for corpus, computer equipment and storage medium | |
| CN111460290B (en) | Information recommendation method, device, equipment and storage medium | |
| CN116663525B (en) | Document auditing method, device, equipment and storage medium | |
| CN110674636A (en) | Power utilization behavior analysis method | |
| CN109800333B (en) | Report generation method and device based on machine learning and computer equipment | |
| CN111583911A (en) | Speech recognition method, device, terminal and medium based on label smoothing | |
| CN111382572A (en) | Named entity identification method, device, equipment and medium | |
| CN113642652A (en) | Method, device and equipment for generating fusion model | |
| CN111651264A (en) | Method and device for acquiring physical machine resource allocation model and computer equipment | |
| CN111552811A (en) | Method and device for information completion in knowledge graph, computer equipment and storage medium | |
| CN111177500A (en) | Data object classification method and device, computer equipment and storage medium | |
| CN118467715B (en) | Method, device, equipment and medium for determining associated equipment | |
| CN113268564B (en) | Method, device, equipment and storage medium for generating similar problems | |
| CN111552810B (en) | Entity extraction and classification method, entity extraction and classification device, computer equipment and storage medium | |
| CN113806568A (en) | Multimedia resource recommendation method and device, electronic equipment and storage medium | |
| CN115794369A (en) | Memory occupation value prediction method and device, storage medium and terminal | |
| CN113761193A (en) | Log classification method and device, computer equipment and storage medium | |
| CN109657710B (en) | Data screening method and device, server and storage medium | |
| CN111754352A (en) | Method, device, equipment and storage medium for judging correctness of viewpoint statement | |
| CN116955534A (en) | Intelligent complaint work order processing method, intelligent complaint work order processing device, intelligent complaint work order processing equipment and storage medium | |
| CN115688002A (en) | Classification method and device, method and device for training classification model and classification model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |