JP7042770B2 - Information processing equipment, information processing methods, and programs - Google Patents

Description

The present invention relates to an information processing apparatus, an information processing method, and a program.

In fields such as electronic commerce, ranking processing of ranking targets (for example, products) is performed based on an input query, and ranking targets with higher ranking results are presented to users. Related to this, the process of receiving the query word string, the process of searching for multiple various product information entries related to the query word column, and the process of extracting the first core product word from the query column from memory. , The process of extracting the second core product word from various product information entries from the memory, and for each second core product word, each second core product word is in the candidate product word list related to the key product word. The process of checking whether or not it exists, the process of reducing the weighting of the product information entry corresponding to each second core product word when each second core product word exists, and the adjustment of the product information entry. The invention of an online product search method comprising a step of sorting product information entries according to the weighting is disclosed (see Patent Document 1).

Special Table 2015-518220 Gazette

In the conventional technology, ranking processing is performed including elements that are not directly related to the query such as the number of page views. Therefore, for example, a popular ranking target is ranked high even though it is not very relevant to the query. In many cases, the ranking target presented may not fully reflect the user's intention.

The present invention has been made in consideration of such circumstances, and an object of the present invention is to provide an information processing device, an information processing method, and a program capable of selecting a ranking target considered to be closer to a search intention. It is one of.

In one aspect of the present invention, the score of the ranking target is derived by inputting the acquisition unit for acquiring the query input by the user and the feature amount of the ranking target matching the query into the trained model. The trained model includes a ranking processing unit that performs ranking processing based on the score, and the trained model is learned by machine learning using the feature amount of the ranking target as training data and the label as teacher data, and the ranking. A category is assigned to each of the targets, and the label indicates a ranking target in which a predetermined action is performed by the user and a ranking target in which the same category as the ranking target in which the predetermined action is performed is assigned. It is an information processing device that is evaluated higher than a ranking target in which the same category as the ranking target in which the predetermined action is not performed by the user is not given.

According to one aspect of the present invention, it is possible to select a ranking target that is considered to be closer to the search intention.

It is a figure which shows an example of the configuration and use environment of the shopping server 100 which used the information processing apparatus which concerns on embodiment. It is a figure which shows an example of the search input screen IM1. It is a figure which shows an example of the search result display screen IM2. It is a figure which shows an example of the product sales screen IM3. It is an image diagram which shows an example of the contents of the log data 152. It is a figure which shows an example of the processing flow by a ranking processing unit 110. It is a figure which shows an example of the contents of the product data 150. It is a figure which shows an example of the content of the processing by a learning processing unit 130. It is a figure for demonstrating the label determination method by a learning processing unit 130. It is a figure which shows the implementation example of the information processing apparatus which concerns on other embodiment.

Hereinafter, embodiments of the information processing apparatus, information processing method, and program of the present invention will be described with reference to the drawings. The information processing device is realized by one or more processors. The information processing device is a device that ranks ranking targets according to a query input by a user. Ranking targets are products (including services) traded in electronic commerce, articles provided via a network, etc., and if a category that is an objective index is given as meta information, It can be anything. Hereinafter, the products traded in electronic commerce will be described as being targeted for ranking.

E-commerce includes shopping, auctions, flea markets, etc. that take place over networks. In the following description, shopping will be described as an example. The information processing device applied to shopping may be a device included in a shopping server or the like that manages a shopping site. That is, the information processing device may be a virtual device. Further, the information processing device may be the shopping server itself. Hereinafter, the information processing apparatus will be described as forming a part of the shopping server. The shopping server may provide a shopping site in response to a request from a browser, or may provide the same content as a shopping site in response to a request from an application program installed on a smartphone or the like. You may. In the following description, the shopping server is assumed to be the former.

[Constitution]
FIG. 1 is a diagram showing an example of a configuration and a usage environment of a shopping server 100 using the information processing device according to the embodiment. The shopping server 100 communicates with the terminal device 10 via the network NW. The network NW includes, for example, the Internet, a WAN (Wide Area Network), a LAN (Local Area Network), a radio base station, a provider terminal, a dedicated line, and the like.

The terminal device 10 is, for example, a mobile phone such as a personal computer or a smartphone, a tablet terminal, or the like. In these, the UA (User Agent) 20 such as a browser or an application program is activated. The UA 20 displays the sales screen provided by the shopping server 100, and also transmits a request corresponding to the input operation made by the user of the terminal device 10 to the shopping server 100.

The shopping server 100 includes, for example, a site providing unit 102, a log collecting unit 104, a ranking processing unit 110, and a learning processing unit 130. The ranking processing unit 110 includes, for example, a search execution unit 112, a feature amount acquisition unit 114, and a model execution unit 116. These components are realized by, for example, one or more hardware processors such as a CPU (Central Processing Unit) executing a program (software). Some or all of these components are hardware (circuit parts) such as LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), GPU (Graphics Processing Unit). It may be realized by (including circuits), or it may be realized by the cooperation of software and hardware. The program may be stored in advance in a storage device (a storage device including a non-transient storage medium) such as an HDD (Hard Disk Drive) or a flash memory, or a removable storage such as a DVD or a CD-ROM. It is stored in a medium (non-transient storage medium) and may be installed by mounting the storage medium in a drive device.

In the present embodiment, an example of the information processing apparatus includes at least a site providing unit 102 and a ranking processing unit 110 that function as a query acquisition unit.

Further, the shopping server 100 stores data such as the trained model 118, the product data 150, and the log data 152 in a storage device such as an HDD, a flash memory, or a RAM (Random Access Memory). This storage device may be attached to the shopping server 100, or may be NAS (Network Attached Storage) that the shopping server 100 can access via the network NW.

The site providing unit 102 provides various screens as a shopping site to the terminal device 10. FIG. 2 is a diagram showing an example of the search input screen IM1. The search input screen IM1 is provided with a query input field A1 for inputting a query for searching for a product, and a search button B1 for executing a search with the input query. When a query is input to the query input field A1 and the search button B1 is operated, the screen transitions to the search result display screen IM2. At this time, the site providing unit 102 acquires a query in the form of form input (functions as a query acquisition unit), outputs the query to the ranking processing unit 110, and requests ranking processing. When a category is selected instead of a query, the screen transitions to the search result display screen IM2 in the same manner.

FIG. 3 is a diagram showing an example of the search result display screen IM2. The search result display screen IM2 includes a search result display field A2. In the search result display column A2, images and descriptions of a predetermined number of products (or all as much as possible) that can be visually recognized by scrolling in the ranking order determined by the ranking processing unit 110 (for example, titles described later). Part or all) are displayed side by side in order of position. Hereinafter, the predetermined number is referred to as k, and the displayed product is referred to as an introduction product. Positions are defined such that the top position is 1, the second position is 2, and so on. When an image or description of one product is operated (selected) on the search result display screen IM2, the screen transitions to the product sales screen IM3.

FIG. 4 is a diagram showing an example of the product sales screen IM3. The product sales screen IM3 includes a product image display column A3-1, a title column A3-2, a detailed explanation column A3-3, and the like. In the title column A3-2 and the detailed explanation column A3-3, the manufacturer of the selected product (selected product), the material of the product, the use, and other contents are posted as text. The text displayed in the title column A3-2 is referred to as a "title", and the text displayed in the detailed explanation column A3-3 is referred to as a "detailed explanation". The product is purchased by the purchaser by operating the product sales screen IM3.

The functions of each part will be described below. The log collection unit 104 collects the usage history of the shopping site for each user and registers it in the log data 152. FIG. 5 is an image diagram showing an example of the contents of the log data 152. As shown in the figure, the log data 152 includes various screens viewed by the user, an input query, a displayed introduction product, a purchased product, and the like for the user ID which is the identification information of the user. Information is associated with it. The log data 152 does not have to be organized by user ID. Further, in reality, the introduced product and the product information may be represented by the product ID described below.

[Ranking process]
FIG. 6 is a diagram showing an example of a processing flow by the ranking processing unit 110. First, the search execution unit 112 of the ranking processing unit 110 searches for the product data 150 by using the query passed from the site providing unit 102. FIG. 7 is a diagram showing an example of the contents of the product data 150. In the product data 150, for example, the product ID, which is the product identification information, is associated with the category ID, which is the category identification information, the product image, the title, the detailed description, the price, and other data. Of these, the product ID is information numbered by the exhibition reception unit (not shown), and the other information is input by the seller. The categories are set hierarchically, for example, and in the illustrated example, for example, "3" indicates "food and drink", "35" indicates "liquor", and "356" indicates "red wine". In the learning process, the learning process unit 130, which will be described later, pays attention to an arbitrary layer of the category (category ID), for example, and performs the process. As an example, the learning processing unit 130 pays attention to the information in the lowest layer (“356” in FIG. 5) and performs processing.

For example, the search execution unit 112 extracts the record of the product data 150 including the query in the title or the detailed description, and acquires the product ID. As a result, the search execution unit 112 acquires the product that matches the query. The search execution unit 112 may acquire the product ID of the product that matches the query by any method. In the example of FIG. 6, it is assumed that n product IDs have been acquired. n does not have to be a fixed value, and the product ID corresponding to the extraction condition may be acquired.

The feature amount acquisition unit 114 acquires the feature amount used as input data by the model execution unit 116 for each product acquired by the search execution unit 112 (for each product ID). The feature amount includes, for example, the number of PV (Page View), the degree of matching with the query, the number of clicks (number of selections), the price, the number of reviews, and the like. The feature amount acquisition unit 114 performs necessary processing on the log data 152, for example, to generate these data.

The feature amount acquisition unit 114 counts, for example, the number of times the product is displayed on the search result display screen in the log data 152 as the number of PVs. Further, the feature amount acquisition unit 114 derives the matching degree with the query based on the character-based matching degree such as the number of matchings between the word string in the query and the word string in the title or the detailed description. Regarding other matters, the feature amount acquisition unit 114 refers to the log data 152 and the product data 150, and performs processing as necessary to acquire the feature amount. Then, the feature amount acquisition unit 114, for example, quantifies and combines various feature amounts to generate one product vector corresponding to the product.

The model execution unit 116 derives a score used for ranking processing by sequentially inputting each of the product vectors generated by the feature quantity acquisition unit 114 into the trained model 118, for example. Then, the ranking processing unit 110 selects, for example, k product IDs in descending order of score and outputs them to the site providing unit 102. The site providing unit 102 searches for product data 150 using k product IDs acquired from the ranking processing unit 110, includes the extracted product information in the search result display screen, and displays it on the user's terminal device 10. Let me. By such processing, it is possible to select a ranking target that is considered to be closer to the search intention and provide information to the user.

[Learning process]
Hereinafter, the generation process of the trained model 118 that brings about the above effects will be described. The learning processing unit 130 performs machine learning using a plurality of data sets in which the feature amount of the product to be ranked is the learning data and the label is the teacher data, and the trained model 118 is generated. The feature amount is the same as that used by the model execution unit 116 as input data, and the training data may be the product vector described above. Therefore, the learning processing unit 130 may acquire the processing result from the feature amount acquisition unit 114 and perform the processing. Although the learning processing unit 130 is included in the shopping server 100 and the information processing device in the present embodiment, it may be configured as a separate information processing device (learning device).

FIG. 8 is a diagram showing an example of the content of processing by the learning processing unit 130. As shown in the figure, the learning processing unit 130 adjusts the parameters of the machine learning model so that the score output by inputting the product vector into the machine learning model becomes higher as the label becomes larger. The machine learning model is, for example, a model constructed based on a neural network, but the model is not limited to this, and any model can be used.

FIG. 9 is a diagram for explaining a label determination method by the learning processing unit 130. For example, as shown in the upper figure of FIG. 9, the introduced products displayed on a certain search result display screen are (1) to (4), and the product (2) is selected by the user to transition to the product sales screen. However, it was not purchased. It is assumed that the product (1) is a product to which the same category ID as the product (2) is given. In this case, the learning processing unit 130 assigns the label value 2 to the product (2) and also assigns the label value 1 to the product (1), for example. Further, as shown in the lower figure of FIG. 9, the introduced products displayed on another search result display screen are similarly (1) to (4), and the product (3) is selected and purchased. It is assumed that the product (4) is a product to which the same category ID as the product (3) is given. In this case, the learning processing unit 130 assigns the label value 4 to the product (3) and assigns the label value 1 to the product (1), for example. In the latter case, the product (1) may be given an arbitrary label value larger than 1 and smaller than 4. The numerical values such as 1, 2, and 4 are merely examples, and the label value given to the product (2) in the example of the upper figure of FIG. 9 is equal to or higher than the label value given to the product (1). It suffices if the relationship that the label value given to (1) is equal to or higher than the label value given to the product (3) and (4) is established, and the label value given to the product (3) in the example of the lower figure of FIG. Is larger than the label value given to the product (4), and the label value given to the product (4) is larger than the label value given to the products (1) and (2). You may use a number. Further, in the above, it is determined that "the same category is given" by focusing on the information of the lowest layer among the hierarchical categories, but when the categories do not match in the lowest layer but the categories match in the upper layer. , A label value smaller than when the categories of the bottom layer match (for example, 0.5) may be given.

In this way, the learning processing unit 130 assigns a label value indicating a high evaluation to the product selected or purchased by the user (the predetermined action is performed), and also assigns a label value indicating a high evaluation to the product in the same category as the selected or purchased product. Label values indicating the next highest evaluation are also given to products. The next highest rated label value is a label value that is lower than the selected or purchased product and higher than the product that is not selected or purchased and is in a different category from the selected or purchased product. Therefore, the label value is given to the selected or purchased product and the product to which the same category as the selected or purchased product is given, and the same category as the selected or purchased product which has not been selected or purchased. It will be evaluated higher than the ranking target that has not been done.

The learning processing unit 130 performs the same processing on the selection or purchase results related to the plurality of search result display screens, assigns a label value to each product, and uses the label as teacher data.

When the label of each product is obtained as described above, the learning processing unit 130 adjusts the parameters of the machine learning model so that the score output by the machine learning model becomes higher as the label is larger. For example, when the product vector of the product (m) is expressed as Vm, the label is expressed as Lm, and the machine learning model is expressed as the function f (Vm), the magnitude relation of the labels is L2> L1> L3> L4. , The learning processing unit 130 adjusts the parameters of the machine learning model in the direction of f (V2)> f (V1)> f (V3)> f (V4). The machine learning model at the time when all the data sets prepared in advance are processed and the adjustment is completed is used as the trained model 118.

By generating the trained model 118 as described above, it is possible to select a product that is considered to be closer to the search intention. In the conventional method, the label is set by evaluating only the selected or purchased product. In this case, high labels will be set for popular products and top-selling products, and it is natural that low-priced products will be selected and purchased more often, so labels for high-priced products will be set. Tended to be low. For this reason, when a user inputs a query with the intention of a less famous product, or when a user inputs a query with the intention of a high-priced product, products different from the search intention may be listed. ..

On the other hand, in the method of the embodiment, even for products in the same category as the selected or purchased product, the label is set so as to give a certain degree of evaluation, so that the product considered to be closer to the search intention is selected. Can be done. This is because the selected or purchased product has a high probability of being in line with the search intent, and the fact that the same category is given is considered to have the same probability of being in line with the search intent. ..

According to the embodiment described above, the score for each product is obtained by inputting the feature amount (product vector) of the product that matches the query into the trained model 118 that outputs the score when the feature amount (product vector) is input. The trained model 118 includes a ranking processing unit 110 that acquires, performs ranking processing based on the score, and selects a ranking target having a higher ranking result, and the trained model 118 is selected or purchased as well as the selected or purchased product. Since the products to which the same category as the products are given are also generated by using a label that gives a high evaluation, it is possible to select a product that is considered to be closer to the search intention.

<Other embodiments>
The information processing device may be configured by a processor separate from the shopping server. FIG. 10 is a diagram showing an implementation example of an information processing apparatus according to another embodiment. As shown in the figure, the information processing device (learning device) A including the learning processing unit 130 and the information processing device (ranking processing device) B including the ranking processing unit 110 are configured separately from the shopping server C, respectively. They may be configured to communicate with each other via the network NW. In this example, the information processing device A and the information processing device B may be integrated into one device. The components having the same reference numerals as those in FIG. 10 in FIG. 10 shall have the same functions as those described with reference to FIG.

In the implementation example of FIG. 10, the learning processing unit 130 of the information processing apparatus A acquires the product data 150 and the log data 152 from the shopping server C at a desired timing, and performs the same processing as described above. Then, the generated trained model is provided to the information processing apparatus B.

The information processing device B includes a query acquisition unit 106. The query acquisition unit 105 acquires a query together with a ranking request from the shopping server C, and passes the query to the ranking processing unit 110. The ranking processing unit 110 uses the product data 150 and the log data 152, which are previously acquired from the shopping server C about once a day, and the learned model 118, which is acquired from the information processing apparatus A, to explain the above. Perform the same processing as the one you did. Also with such a configuration, it is possible to select a product that is considered to be closer to the search intention, as in the above embodiment.

Although the embodiments for carrying out the present invention have been described above using the embodiments, the present invention is not limited to these embodiments, and various modifications and substitutions are made without departing from the gist of the present invention. Can be added.

100 Shopping server (information processing device)
102 Site provision unit 104 Log collection unit 106 Query acquisition unit 110 Ranking processing unit 112 Search execution unit 114 Feature quantity acquisition unit 116 Model execution unit 118 Learned model 130 Learning processing unit 150 Product data 152 Log data A, B Information processing device C Shopping server

Claims (7)

An acquisition unit that acquires the query entered by the user,
It is provided with a ranking processing unit that derives the score of the ranking target by inputting the feature amount of the ranking target that matches the query into the trained model and performs ranking processing based on the score.
The trained model is trained by machine learning using the features of the ranking target as training data and labels as teacher data.
A category is assigned to each of the ranking targets.
The label is a ranking target in which a predetermined action is performed by the user, and a ranking target to which the same category as the ranking target in which the predetermined action is performed is given, and the user has not performed the predetermined action. It is evaluated higher than the ranking target to which the same category as the ranking target in which the predetermined action is performed is not assigned.
Information processing equipment. The predetermined action is an action of selecting the presented ranking target by using the terminal device of the user.
The information processing apparatus according to claim 1. The label evaluates a ranking target in which a predetermined action is performed by a user higher than a ranking target in which the same category as the ranking target in which the predetermined action is performed is given.
The information processing apparatus according to claim 1 or 2. The ranking target in which the predetermined action is performed by the user and the ranking target to which the same category as the ranking target in which the predetermined action is performed are given, the predetermined action is not performed by the user and the predetermined action is performed. Generates a label that is evaluated higher than the ranking target that is not given the same category as the ranking target, and generates the trained model by machine learning using the feature amount of the ranking target as training data and the label as teacher data. Further equipped with a learning processing unit
The information processing apparatus according to any one of claims 1 to 3. Obtain log data showing user behavior and data on multiple ranking targets with categories assigned to each.
The ranking target in which the predetermined action is performed by the user and the ranking target to which the same category as the ranking target in which the predetermined action is performed are given, the predetermined action is not performed by the user and the predetermined action is performed. Generates a label that is rated higher than the ranking target that is not given the same category as the ranking target
It is provided with a learning processing unit that generates a trained model by machine learning using the feature amount of the ranking target as training data and the label as teacher data.
Information processing equipment. The computer
Get the query entered by the user and
It is an information processing method that derives the score of the ranking target by inputting the feature amount of the ranking target that matches the query into the trained model and performs the ranking process based on the score.
The trained model is trained by machine learning using the features of the ranking target as training data and labels as teacher data.
A category is assigned to each of the ranking targets.
The label is a ranking target in which a predetermined action is performed by the user, and a ranking target to which the same category as the ranking target in which the predetermined action is performed is given, and the user has not performed the predetermined action. It is evaluated higher than the ranking target to which the same category as the ranking target in which the predetermined action is performed is not assigned.
Information processing method. On the computer
Get the query entered by the user
It is a program that derives the score of the ranking target by inputting the feature amount of the ranking target that matches the query into the trained model, and causes the ranking process to be performed based on the score.
The trained model is trained by machine learning using the features of the ranking target as training data and labels as teacher data.
A category is assigned to each of the ranking targets.
The label is a ranking target in which a predetermined action is performed by the user, and a ranking target to which the same category as the ranking target in which the predetermined action is performed is given, and the user has not performed the predetermined action. It is evaluated higher than the ranking target to which the same category as the ranking target in which the predetermined action is performed is not assigned.
program.

Images