JP6501936B1 - INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND PROGRAM - Google Patents

Abstract

An information processing apparatus, an information processing method, and a program capable of relatively widely extracting the number of users who may be interested in a given keyword. A first extraction unit extracts a second keyword that is likely to co-occur with a first keyword in the action history of each of a plurality of users on the network, and the network includes the first keyword or the second keyword. A second extraction unit for extracting a user having an action history, and one or more users whose action history on the network is similar to the user extracted by the second extraction unit from among a plurality of users; An information processing apparatus comprising: a third extraction unit that extracts a user corresponding to a keyword. [Selected figure] Figure 1

Description

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

  Based on the search keyword information extraction means for extracting search keyword information contained in the search history information of each user using the information search engine, and based on the extracted search keyword information, a class indicating the trend of the input keyword of each user is generated There is known a user classification device including class generation means, classification means for classifying each user in the generated class, and classification result presentation means for presenting at least classified classification results (see Patent Document 1).

  The apparatus further includes core keyword extraction means for extracting core keywords representing classes generated by the class generation means, and the classification result presentation means presents the extracted core keywords together with the user classification results. The core keyword extraction unit includes, as information for extracting the core keyword, word co-occurrence information in the summary included in the Web search result for each user's input keyword and word co-occurrence information in the classified class. Extract core keywords by comparison.

JP, 2009-43125, A

  In the prior art, it may not be possible to secure a sufficient number of users having similar properties. For this reason, for example, in order to use it for applications, such as marketing, the user might be narrowed down too much.

  The present invention has been made in consideration of such circumstances, and is an information processing apparatus capable of relatively widely extracting the number of users who may be interested in a given keyword. , An information processing method, and a program are provided.

  One aspect of the present invention includes a first extraction unit that extracts a second keyword that is likely to co-occur with a first keyword in the behavior history of each of a plurality of users, and the first keyword or the second keyword. A second extraction unit for extracting a user having an action history on the network; and one or more users whose action history on the network is similar to the user extracted by the second extraction unit among the plurality of users; It is an information processor provided with the 3rd extraction part extracted as a user corresponding to the 1st keyword.

  According to one aspect of the present invention, the number of users who may be interested in a given keyword can be extracted relatively widely.

It is a figure which shows an example of a structure of service server 100 using an information processing apparatus, and use environment. It is a figure which shows an example of the content of the user information 194. FIG. It is a figure which shows an example of the content of the search log 196. FIG. It is a figure for demonstrating the process of the learning model production | generation part 154. FIG. It is a flowchart which shows an example of the flow of the process performed by an information processing apparatus. It is a figure which shows an example of a structure of the shopping server 200 using an information processing apparatus, and a use environment. It is a figure which shows an example of the content of goods etc. data 292. FIG. It is a figure which shows an example of the content of the purchase log 296. FIG.

  Hereinafter, embodiments of an information processing apparatus, an information processing method, and a program of the present invention will be described with reference to the drawings.

  The information processing apparatus is realized by one or more processors. The information processing apparatus is used for the purpose of relatively widely extracting users (users) who may be interested in a certain keyword (first keyword). For example, when an enterprise wants to offer electronic coupons or advertisements only to users who may be interested in keywords representing a product or service provided by the enterprise, The information processing apparatus of the present invention is used.

  The information processing apparatus extracts a second keyword which is likely to co-occur with the first keyword, extracts a user (primary extracted user) having an action history on the network including the first keyword or the second keyword, and performs primary extracted Machine learning is performed with the user's action history on the network as correct data, and when the target user's action history on the network is input, a learning model is generated that outputs information indicating the similarity with the primary user to be extracted. Since the primary target user is a user who is highly likely to be interested in the first keyword, the user judged by the learning model to be highly similar to the primary target user is interested in the first keyword. It is presumed that there is a possibility of Therefore, this learning model is for outputting information that allows the target user to determine whether or not the user is interested in the first keyword. In addition, as long as it is a method capable of extracting a user whose action history on the network is similar to the primary extracted user, any method may be used without being limited to the method of generating a learning model.

  Here, if only the primary extracted users are extracted, the number of extracted users is often limited, but in the information processing apparatus of the present invention, the learning model is generated to be an object. By expanding the number of users, the number of users who may be interested in the given first keyword can be extracted relatively widely.

  The action history on the network means, for example, the history of a query input for a search, or the history (purchase history) of a product or the like purchased in an electronic commerce for selling a product or service (hereinafter, product or the like) It is. In the first embodiment, the former will be described, and in the second embodiment, the latter will be described.

  The information processing apparatus may be an apparatus that realizes functions alone or may be a virtual apparatus included in an apparatus (web server, application server, etc.) having other functions. In the following description, the information processing apparatus will be described as being included in a service server that provides content to a user's terminal device, and a shopping server that provides a shopping site.

First Embodiment
[Constitution]
FIG. 1 is a diagram showing an example of the configuration and usage environment of a service server 100 using an information processing apparatus. In the illustrated example, one or more terminal devices 10 and the request source server 300 are connected to the service server 100 via the network NW. The network NW includes, for example, the Internet, a wide area network (WAN), a local area network (LAN), a provider terminal, a wireless communication network, a wireless base station, a dedicated line, and the like. The components shown in FIG. 1 include a communication interface for connecting to the network NW and other networks. The communication interface includes a network card such as a network interface card (NIC), a wireless communication module, and the like.

[Terminal device]
The terminal device 10 is, for example, a mobile phone such as a smartphone, a tablet terminal, various personal computers, and the like. In the terminal device 10, a UA (User Agent) such as a browser or an application program is activated, and transmits a request according to the content input by the user to the service server 100. Also, the UA displays various images based on the information acquired from the service server.

[Service server]
The service server 100 is a web server that provides a web page to the terminal device 10 in response to a request from a browser, or an application server that provides the terminal device 10 that provides images and sounds in response to a request from an application program.

  The service provided by the service server 100 is provided as a customized service for more individual users, for example, by inputting a user ID and a password and logging in. When the service server 100 is a web server, it is possible to receive provision of a web page without logging in, but in that case, a general-purpose web page which is not customized for each user is provided.

  For example, the service server 100 includes a content providing unit 110, a user management unit 120, a search execution unit 130, a first extraction unit 150, a second extraction unit 152, a learning model generation unit 154, and a third extraction unit. 156 and the privilege giving unit 170. Some or all of these components are hardware (circuit units) such as LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), GPU (Graphics Processing Unit), etc. may be realized by the cooperation of software and hardware. The program may be stored in advance in a storage device such as a hard disk drive (HDD) or a flash memory, or is stored in a removable storage medium such as a DVD or a CD-ROM, and the storage medium is stored in the drive device. It may be installed by being attached. The “information processing apparatus” includes at least a first extraction unit 150, a second extraction unit 152, a learning model generation unit 154, and a third extraction unit 156 among the components.

  The service server 100 may also include a storage unit 190. The storage unit 190 may be an external storage device such as a NAS (Network Attached Storage) accessible by the service server 100 via the network NW. The storage unit 190 stores information such as content data 192, user information 194, and search log 196.

  The content providing unit 110 provides the terminal device 10 with content based on the content data 192. The content data 192 is, for example, a news article, a moving image, a still image, a voice, or the like, or reference information (eg, URL; Uniform Resource Locator) for referring to them. The content provided by the content providing unit 110 has a function of inputting a query and instructing a search. The content providing unit 110 passes the query input in the terminal device 10 to the search execution unit 130 to execute the search, and provides the terminal device 10 with a screen indicating the search result. The query may include a suggestion query or the like in addition to the query directly input.

  The user management unit 120 manages a user who logs in to a service provided by the service server 100 (content provision in this example) by using the user information 194. FIG. 2 is a diagram showing an example of the content of the user information 194. As shown in FIG. The user information 194 is, for example, information in which age, gender, birthday, occupation, and other information are associated with a user ID that is identification information of the user.

  The search execution unit 130 executes a search on the network NW in response to an instruction from the content providing unit 110. Detailed techniques for searching in the network will not be described because various techniques have already been disclosed. The search execution unit 130 registers the input query in the search log 196 each time a search is performed. FIG. 3 is a diagram showing an example of the content of the search log 196. As shown in FIG. The search log 196 is, for example, information in which a query is associated with a search time for each user ID. Hereinafter, a set of queries associated with a user ID of a certain user may be referred to as “user's query history”. The search log 196 may be added with information indicating the break of the session. A session is, for example, a valid period of a state management function such as a cookie. For example, a period from access to a web page in a website to a predetermined time (timeout) is treated as one session. Also, a session may be a period from accessing a certain web page in a web site to switching to another web page in the web site or a web page in another web site. It may be a period from accessing a web page in the site to closing a web browser displaying the web page. Further, the browsing log may be stored separately from the search log 196, and the search log 196 may be held in the storage unit 190 so as to be included in the browsing log.

(Information processing device)
Hereinafter, components constituting the information processing apparatus will be described. The processing of each component in the information processing apparatus may be performed without setting a time limit, or may be performed with a time limit such as one year, several months, one month, or within the range of a session.

  The first extraction unit 150 acquires the first keyword from the request source server 300. The business operator operating the request source server 300 requests extraction of a user by designating a keyword for the purpose of “I want to give a privilege such as an electronic coupon to a user who may be interested in this keyword”. . The first keyword is a keyword related to this request.

  The first extraction unit 150 extracts a second keyword that is likely to co-occur with the first keyword (for example, has a high probability) in the search log 196 that is an example of the behavior history of each of the plurality of users on the network. Although an example of processing which extracts the 2nd keyword which tends to co-occur with the 1st keyword is explained below, if it is the method which can extract the 2nd keyword not just this one example with similar tendency, any method You may use. For example, the first extraction unit 150 sequentially selects candidates for the second keyword from all the queries in principle included in the search log 196 (the number may be excluded). The population of candidates for the second keyword is not limited to the query included in the search log 196, but may be a keyword included in any dictionary or website.

The first extraction unit 150 obtains, for all users, the ratio of users in which both the first and second keywords are included in the query history (co-occurring user ratio PP). The co-occurrence user ratio PP is, for example, expressed by the conditional probability formula (1). In addition, the first extraction unit 150 determines, for all users, the proportion of users who do not include the first keyword in the query history and the second keyword in the query history (non-co-occurrence user ratio PN). The non-co-occurrence user ratio PN is, for example, expressed by the conditional probability formula (2).
PP = P (first query | second query) ... (1)
Pn = P (first query | not second query) ... (2)

  Then, based on the difference between the co-occurrence user ratio PP and the non-co-occurrence user ratio PN, the first extraction unit 150 extracts the second keyword having a high probability of co-occurrence with the first keyword. For example, the first extraction unit 150 calculates the difference between the co-occurring user ratio PP minus the non-co-occurring user ratio PN or the calculation result such as the quotient obtained by dividing the co-occurring user ratio PP by the non-co-occurring user ratio PN. The second keyword which is higher in the above or the whole of the second keyword candidates is extracted as a keyword having a high probability of co-occurring with the first keyword. There is no particular limitation on the number of queries extracted as the second keyword, and the first extraction unit 150 may extract an arbitrary number of queries as the second keyword.

  The second extraction unit 152 extracts a user in which at least one of the first keyword and the second keyword is included in the query history. It is assumed that “baseball” is designated as the first keyword among the queries illustrated in FIG. 3 and “hawks” is extracted as one of the second keywords. In this case, the second extraction unit 152 extracts a set of users who have input “baseball” or “hawks” as a query. 3, the second extraction unit 152 extracts the user whose user ID is "AAA", the user who is "BBB", and the user whose user ID is "CCC". The called user is referred to as "primary user to be extracted".

  The learning model generation unit 154 performs machine learning using the feature quantities of the primary extracted user as correct data, and a learning model for extracting a user having a feature quantity close to that of the primary extracted user (hereinafter, secondary extracted user). Generate

  FIG. 4 is a diagram for explaining the process of the learning model generation unit 154. The feature quantity handled by the learning model generation unit 154 is, for example, a vector (for example, 10,000-dimensional vector having 10,000 elements) in which each of a plurality of queries similar to the population of the second keyword is associated with the element This is a vector having 1 if the query corresponding to each element is a query (not limited to the first keyword or the second keyword) input by the user in the past, and zero otherwise. This vector may be referred to as a sparse vector because the queries the user enters are not exhaustive. Hereinafter, this is referred to as a user vector. FIG. 4 exemplifies the user vector of the user whose user ID in FIG. 3 is “DDD”.

  The learning model generation unit 154 generates some learning model by machine learning such that the score value becomes high when the user vector of the primary extracted user is input. The learning model is generated using, for example, a neural network such as DNN (Deep Neural Network), an activation function, and the like as a component. Specifically, the learning model is a data structure (software) including node information and weight values in the neural network, parameters of the activation function, and the like.

  When such a learning model is generated, even if a user who is not the primary extracted user and has a query history similar to that of the primary extracted user is input, the user vector of the primary extracted user As in the case of entering, a high score value will be obtained. Since the secondary extracted user who obtained this high score value is a user having a user vector close to the primary extracted user in the vector space, the primary extracted user is interested in the first keyword which is the key to be extracted It is inferred that the user is likely to have

  Temporarily, it is assumed that the tendency "a user who is interested in baseball likes ramen" exists in the world. In this case, even if "ramen" is not extracted as the second keyword, the query history of the primary user to be extracted includes queries such as "ramen", "chicken galley soup", "delicious noodles", and "Chinese buckwheat noodles" The possibility is high. The learning model is not only the first keyword and the second keyword, but also those tendencies learned together. As a result, when the user vector of the user whose query regarding the frame is included in the query history is input to the learning model, a high score value is obtained. The user with the user ID "DDD" illustrated in FIG. 4 is extracted as a secondary user who may be interested in "baseball" although "baseball" and "hawks" are not included in the query history. there is a possibility. As described above, when the information processing apparatus extracts the number of users who may be interested in the given first keyword, for example, the user who simply includes the first keyword and the second keyword in the query history. Can be widely extracted.

  The third extraction unit 156 extracts one or more users whose high score values are output by the learning model as secondary extracted users.

  The privilege imparting unit 170 imparts privileges such as an electronic coupon and a lottery right to the primary user to be extracted and the secondary user to be extracted. The content of the granted benefit is notified to the logged-in user among the content provided by the content providing unit 110. A user who has been given a benefit can exercise the benefit in various situations.

(Processing flow)
FIG. 5 is a flowchart showing an example of the flow of processing executed by the information processing apparatus. First, the first extraction unit 150 acquires the first keyword from the request source server 300 (S100). The first extraction unit 150 refers to the search log 196 (S102), and calculates the co-occurrence user ratio PP and the non-co-occurrence user ratio PN for each second keyword candidate (for example, a query included in the search log 196). (S104). Then, the first extraction unit 150 extracts, as a second keyword, a query in which the co-occurrence user ratio PP is larger than the non-co-occurrence user ratio PN as a second keyword (S106).

  Next, the second extraction unit 152 extracts a user including the first keyword or the second keyword in the query history as a primary extracted user (S108).

  Next, the learning model generation unit 154 generates a user vector for each user whose user ID is included in the user information 194 (S110). Then, the learning model generation unit 154 performs machine learning so that the score value of the primary extracted user becomes high, and generates a learning model (S112).

  Next, the third extraction unit 156 (which may be another functional unit) inputs a user vector to the learning model for each user whose user ID is included in the user information 194 (S114). Then, the third extraction unit 156 extracts a user related to a user vector with a high score value as a secondary extracted user (S116). The information processing apparatus outputs the primary extracted user, the secondary extracted user, and the user who may be interested in the first keyword (S118).

  According to the first embodiment described above, the second keyword that is likely to co-occur with the first keyword is extracted from the query history (action history on the network) of each of a plurality of users, and the first keyword or the second keyword is extracted. A user having a query history (activity history on the network) including a keyword is extracted as a primary extracted user, machine learning is performed with the user vector (feature amount) of the primary extracted user as correct data, and the target user's user When a vector (feature amount) is input, a learning model that outputs information indicating the similarity to the primary extracted user is generated, and among a plurality of users, one or more users whose query history is similar to the primary extracted user May be interested in the given first keyword by extracting as a user corresponding to the first keyword The number of a user, can be relatively widely extraction.

Second Embodiment
The second embodiment will be described below. In the first embodiment, the action history on the network is the query history. In the second embodiment, the action history on the network is a history of the name or the like of a product or service purchased in an electronic commerce via a shopping site or an auction site. In the following description, the information processing apparatus is described as being included in a shopping server that provides a shopping site.

[Constitution]
FIG. 6 is a diagram showing an example of the configuration and usage environment of the shopping server 200 using the information processing apparatus. The shopping server 200 is a web server that provides a web page to the terminal device 10 in response to a request from a browser, or an application server that provides the terminal device 10 that provides images and sounds in response to a request from an application program.

  The service provided by the shopping server 200 is provided, for example, by inputting a user ID and a password and logging in.

  In the configuration shown in FIG. 6, the functions of the content providing unit 210 and the user management unit 220 are the same as the functions of the content providing unit 110 and the user management unit 120 in the first embodiment, and thus the description thereof is omitted. Also, the user information 294 may be the same information as the user information 194 in the first embodiment.

  The sales management unit 230 refers to the product etc. data 292 and configures a product etc. introduction column embedded in the shopping site provided by the content providing unit 210. FIG. 7 is a diagram showing an example of the contents of the product etc. data 292. As shown in FIG. The product etc. data 292 is information in which items such as categories, titles, and detailed descriptions hierarchically represented are corresponded to a product etc. ID which is identification information of a product etc., for example.

  In addition, the sales management unit 230 registers information on purchased products and the like in the purchase log 296 each time a product or the like is purchased on the shopping site. FIG. 8 is a diagram showing an example of the contents of the purchase log 296. As shown in FIG. The purchase log 296 is, for example, information in which information such as a purchased product is associated with a search time for each user ID. The browsing log may be stored separately from the purchase log 296, and the purchase log 296 may be held in the storage unit 290 as being included in the browsing log.

The information processing apparatus according to the second embodiment sets a keyword extracted from information such as a product registered in the purchase log 296 as a candidate for the second keyword, instead of the query. The keywords extracted from the information of goods etc. registered in the purchase log 296 are, for example, information of the lowermost layer of the category, proper nouns extracted from the title by morphological analysis, and the like. The extracted keywords for each user are referred to as a purchase history.

  The other aspects are the same as in the first embodiment, and thus detailed description will be omitted. Although the first extraction unit 250, the second extraction unit 252, the learning model generation unit 254, and the third extraction unit 256 are different in type of input data, the first extraction unit 150, the second extraction unit of the first embodiment The same processing as the learning model generation unit 154 and the third extraction unit 156 is performed. Further, the privilege imparting unit 270, similarly to the privilege imparting unit 170, imparts a privilege such as an electronic coupon or a lottery right to the primary user to be extracted and the secondary user to be extracted.

  According to the second embodiment described above, the second keyword that is likely to co-occur with the first keyword is extracted from the purchase history (action history on the network) of each of the plurality of users, and the first keyword or the second keyword is extracted. A user having a purchase history (activity history on the network) including a keyword is extracted as a primary extracted user, machine learning is performed with the user vector (feature amount) of the primary extracted user as correct data, and the target user's user When a vector (feature amount) is input, a learning model that outputs information indicating the similarity to the primary extracted user is generated, and among a plurality of users, one or more users whose purchase histories are similar to the primary extracted user May be interested in the given first keyword by extracting as a user corresponding to the first keyword The number of over The, can be relatively widely extraction.

<Other Embodiments>
In the second embodiment, the candidate of the second keyword is extracted from the purchase history of the product etc. of the user, but instead, the history of transition from the shopping site or auction site to the product introduction screen, ie, product etc. The candidate of the second keyword may be extracted from the browsing history of the introduction screen. The browsing history is, like the purchase history, information on the lowermost layer of a category such as a product for which the introduction screen is browsed, a proper noun extracted from the title by morphological analysis, and the like.

  Alternatively, instead of the query history, purchase history, and browsing history of the introduction screen such as products, the second keyword candidate may be extracted from the history of the news article browsed by the user. In this case, the information processing apparatus may set a representative noun (for example, a noun having a high tf-idf value) included in the news article as a candidate for the second keyword.

  As mentioned above, although the form for carrying out the present invention was explained using an embodiment, the present invention is not limited at all by such an embodiment, and various modification and substitution within the range which does not deviate from the gist of the present invention Can be added.

10 terminal device 100 service server 110, 210 content providing unit 120, 220 user management unit 130 search execution unit 150, 250 first extraction unit 152, 252 second extraction unit 154, 254 learning model generation unit 156, 256 third extraction unit 170, 270 bonus granting unit 190, 290 storage unit 192 content data 194, 294 user information 196 search log 200 shopping server 230 sales management unit 292 product etc. data 296 purchase log 300 request source server

Claims (8)

A first extraction unit that extracts a second keyword that is likely to co-occur with the first keyword in the behavior history on the network of each of a plurality of users;
A second extraction unit for extracting a user having an action history on the network including the first keyword or the second keyword from among a plurality of users;
A third extraction unit which extracts, from among a plurality of users, one or more users whose action history on the network is similar to the user extracted by the second extraction unit, as a user corresponding to the first keyword;
An information processing apparatus comprising: When machine learning is performed using the feature amount related to the user's activity history on the network extracted by the second extraction unit as correct data, and the feature amount related to the activity history on the target user's network is input, the second extraction unit And a generation unit that generates a learning model that outputs information indicating the similarity to the user extracted by
The third extraction unit extracts, from among a plurality of users, one or more users to which information indicating that the learning model has high similarity to the user extracted by the second extraction unit is output.
An information processing apparatus according to claim 1. The first extraction unit includes, in the action history, a probability that the first keyword is included in the action history of the user whose keyword is a candidate for the second keyword is included in the action history, and a keyword that is a candidate for the second keyword The second keyword is extracted based on the degree of difference with the probability that the first keyword is included in the action history of the user who is not
The information processing apparatus according to claim 1. The feature value is a vector whose element is a value representing whether each of a plurality of keywords related to the action history on the network is included.
The information processing apparatus according to claim 2. The action history on the network is a query history,
The information processing apparatus according to any one of claims 1 to 4 . The action history on the network is a purchase history in electronic commerce,
The information processing apparatus according to any one of claims 1 to 4 . The computer is
Extract a second keyword that is likely to co-occur with the first keyword in the action history on the network of each of a plurality of users,
Extracting a user having an activity history on the network including the first keyword or the second keyword;
Extracting one or more users whose behavior history on the network is similar to the extracted user among the plurality of users as the user corresponding to the first keyword;
Information processing method. On the computer
The second keyword that is likely to co-occur with the first keyword is extracted in the action history on the network of each of a plurality of users,
Causing a user having an activity history on the network including the first keyword or the second keyword to be extracted;
Among the plurality of users, one or more users whose action history on the network is similar to the extracted user are extracted as a user corresponding to the first keyword.
program.

Images