JP2019191686A - Information processing apparatus, information processing system, information processing method, and program - Google Patents

Abstract

To provide an information processing apparatus, an information processing system, an information processing method, and a program capable of accurately extracting a query that a user focuses on periodically.SOLUTION: An information processing apparatus includes: an evaluation unit for evaluating queries based on user behavior which is for service provided in response to the queries entered by a user in a first period and which is managed for each type of service and user behavior which is for service provided in response to the query entered by the user in a second period and which is managed for each type of service; and a determination unit for determining a periodic query having predetermined periodicity based on an evaluation result of the evaluation unit.SELECTED DRAWING: Figure 1

Description

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

  Conventionally, a technique relating to a query determination device that determines whether a predetermined query is a specific query has been disclosed (for example, see Patent Document 1). The query determination device refers to the click log, and totals the total number of clicks for the total number of clicks of links to pages included in the search result list for the predetermined query, and the click log and the predetermined query. Number of clicks on a specific Web page that counts the number of clicks on links to a specific Web page in the search result list for a given query based on a determination result of whether or not a page included in the search result list is a specific Web page A totaling unit, a specific Web page click rate calculation unit that calculates a ratio of the number of clicks of a link to a page determined to be a specific Web page to the total number of clicks, and when this ratio is equal to or greater than a predetermined value, A specific query determining unit that determines that the query is a specific query.

JP 2012-78888 A

  However, in the above-described conventional technique, there is a case where it is not possible to accurately extract a query that a user pays attention to periodically.

  The present invention has been made in consideration of such circumstances, and provides an information processing apparatus, an information processing system, an information processing method, and a program capable of accurately extracting a query that a user pays attention to periodically. One of the purposes is to do.

  One aspect of the present invention is a user behavior for a service provided in response to a query input by a user in the first period, the user behavior managed for each type of the service, and a user in the second period. An evaluation unit that evaluates the query based on a user's behavior for a service provided in response to the query input by the user and is managed for each type of the service; and It is an information processing apparatus provided with the determination part which determines the periodic query which has predetermined | prescribed periodicity based on an evaluation result.

  According to one aspect of the present invention, it is possible to accurately extract a query that a user focuses on periodically. Furthermore, by using a query that the user focuses on periodically and generating a model that determines the type of service according to the query, marketing more appropriately or providing useful information to the user Can do.

It is a figure which shows an example of a function structure of the information processing system 1 containing an information processing apparatus. It is a figure which shows an example of the content of the action history information. It is a figure which shows an example of image IM displayed on the terminal device when providing the specific information of the type of service according to the input query to a user. It is a figure which shows an example of the content of the analysis information 124. It is a figure which shows an example of the content of the score information. 6 is a flowchart illustrating an example of a flow of processing executed by the information processing apparatus 100. It is a figure which shows an example of the periodicity query derived | led-out by the information processing apparatus 100. FIG. 10 is a flowchart illustrating another example of the flow of processing executed by the information processing apparatus 100. It is a figure which shows an example of the content of the learning data. It is a figure which shows an example of the learned model 132. FIG. 10 is a flowchart illustrating another example of the flow of processing executed by the service server 20 and the information processing apparatus 100. It is a figure which shows an example of the images IM1 and IM2 displayed on the terminal device. It is a figure which shows an example of image IM3 which the service server 20 displays on the display part of the terminal device 10 in 2nd Embodiment. It is a figure which shows an example of the analysis information 124A of 2nd Embodiment. It is a figure which shows an example of the content of the analysis information 124B of 3rd Embodiment.

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

[Overview]
The information processing apparatus is realized by one or more processors. The information processing apparatus evaluates a query input by the user, and determines a periodic query having a predetermined periodicity based on the evaluation result. More specifically, the information processing apparatus is a user action for a service provided in response to a query input by the user in the first period, and the user action managed for each type of service; The query is evaluated based on the user's behavior with respect to the service provided in response to the query input by the user during the period, and the user behavior managed for each type of service.

  “Provided in response to a query” means that a service is provided by setting a link on a screen showing a search result based on the input query. Also, “provided in response to a query” means, for example, that a service is included and provided in a predetermined area (shortcut area) of an image provided to a user according to an input query. “Service types” are classified into, for example, electronic commerce, cooking recipes, map information, and the like. “User behavior” means, for example, that a link is selected on a screen showing the result of a query, or that an action corresponding to conversion is taken after the selection. “Select” includes clicks, touches, taps and other modes. In the following description, it is assumed that “click” is performed as an example.

  The information processing apparatus learns, as learning data, data in which a periodic query and a service type are associated with each other, and generates a learned model. This learned model is used for which service type information should be provided to the user when a query is input by the user. The information processing apparatus uses the learned model to determine to provide the user with a service related to the type of service corresponding to the query.

<First Embodiment>
FIG. 1 is a diagram illustrating an example of a functional configuration of an information processing system 1 including an information processing apparatus. The information processing system 1 includes, for example, one or more terminal devices 10, one or more service servers 20, and an information processing device 100. The service server 20 and the information processing apparatus 100 may be integrated. These devices communicate with each other via the network NW. The network NW includes, for example, a WAN (Wide Area Network), a LAN (Local Area Network), the Internet, a dedicated line, a wireless base station, a provider, and the like.

[Terminal device]
The terminal device 10 is a device used by a user, and is, for example, a mobile phone such as a smartphone, a tablet computer, a notebook computer, or a desktop computer. In the terminal device 10, a user agent (UA) such as a browser or an application program is activated and performs the following operations. A user operates the terminal device 10 to input a query and requests the service server 20 to search for information related to the input query. The terminal device 10 acquires information related to the query that is the search result of the service server 20, and the acquired information is displayed on the display unit of the own device. And a user acquires information by selecting the area | region which shows the location of the provision source which provides the information of interest among the information regarding a desired query.

[Service Server]
The service server 20 provides some website having a query input field. Instead of providing a website, an image displayed in an application program may be provided. In the following description, it is assumed that the service server 20 is a web server that provides a website. The service server 20 includes a search unit 22. For example, the service server 20 acquires a query from the terminal device 10 and causes the search unit 22 to search for information related to the acquired query via the network NW. The service server 20 provides the search result of the search unit 22 to the terminal device 10. Further, the service server 20 stores action history information, which is a user action history, in a storage device. The action history information is information in which a query input by a user, a service type, a time, and a user action are associated with each other. For details, see FIG. 2 described later.

[Information processing device]
The information processing apparatus 100 includes, for example, a communication unit 102, an evaluation unit 104, a determination unit 106, a learning unit 108, a processing unit 110, and a storage unit 120. The evaluation unit 104, the determination unit 106, the learning unit 108, and the processing unit 110 are realized by, for example, a hardware processor such as a CPU (Central Processing Unit) executing a program stored in a storage device. In addition, these functional units may be realized by hardware 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 cooperation of software and hardware. Further, the above program may be stored in advance in a storage device, or stored in a removable storage medium such as a DVD or a CD-ROM, and the storage medium is attached to the drive device of the information processing apparatus 100. May be installed in the storage device.

  The storage unit 120 is realized by, for example, a ROM (Read Only Memory), a flash memory, an SD card, a RAM (Random Access Memory), a HDD (Hard Disc Drive), a register, and the like. Further, a part or all of the storage unit 120 may be a NAS (Network Attached Storage), an external storage server device, or the like. The storage unit 120 stores, for example, action history information 122, analysis information 124, score information 126, periodic query information 128, learning data 130, and a learned model 132. Details of each information will be described later.

  The communication unit 102 includes a communication interface such as a network interface card.

  The evaluation unit 104 is a user behavior for a service provided in response to a query input by the user in the first period and is managed for each service type, and a user input in the second period. The query is evaluated based on the user behavior for the service provided in response to the query and the user behavior managed for each type of service.

  The determination unit 106 determines a periodic query having a predetermined periodicity based on the evaluation result of the evaluation unit 104.

  The learning unit 108 uses the data in which the periodic query and the service type are associated with each other as the learning data, and when the query is input, the learned model 132 that is used to provide the service type information to the user. Is generated. Further, the learning data further includes data in which the query input by the user and the service type are associated with each other during the most recent period (for example, the most recent months) when the learned model 132 is generated.

  The processing unit 110 inputs a query input by the user to the learned model 132, and determines the type of service provided to the user based on the output result output by the learned model 132.

[Behavior history information]
FIG. 2 is a diagram illustrating an example of the contents of the action history information 122. The action history information 122 is action history information transmitted from the service server 20. In the example of FIG. 2, the query input by the user, the type of service related to the specific information provided to the user, the information indicating whether or not the specific information has been clicked, and the year and month are associated with each other. . The specific information is information of a predetermined service type provided when the user inputs a query. The specific information is, for example, information (information related to the service) included in the shortcut area (see area AR1 in FIG. 3). When the specific information is clicked, a transition is made to a link destination associated with the specific information, for example, a shopping web page, a cooking recipe web page, a map web page, or the like.

  Hereinafter, an example of the image IM displayed on the terminal device 10 when the service server 20 provides the user with specific information of the type of service corresponding to the input query will be described with reference to FIG. For example, when a query is input, the service server 20 uses the evaluation learned model (or a preset reference) to determine the type of service according to the query.

  For example, the learned model for evaluation is a model learned by using, as teacher data, a query in which the type of service of specific information included in the area AR1 is known. The teacher data may include a query (a negative example) in which the type of service of specific information not included in the area AR1 is known. The learned model for evaluation is generated using, for example, a neural network (Neural Network) or a support vector machine (SVM) as a component. In the following description, it is assumed that the service type is determined to be “recipe” when the query “Nanakusa” is input to the learned model for evaluation in a predetermined period. Note that when “Nanagusa” is input, the service type may be determined as “recipe”, “shopping”, or “map” according to a predetermined standard or uniformly.

  The service server 20 includes information (location, image, etc.) related to a web page that provides a method of making a herb based on the query “Nanagusa” and the determined service type “recipe”, and the query “Nanagusa” Get other information about (for example, information related to encyclopedias that describe the herbs). Then, the service server 20 generates information for displaying the image IM on the display unit of the terminal device 10 based on the acquired information, and transmits the generated information to the terminal device 10.

  In the area IM1 of the image IM, for example, an image indicating a link destination of a web page that provides a method of making a herb rice that is a service type “recipe” determined based on the learned model for evaluation is displayed. In the area IM2 of the image IM, for example, information indicating a link destination of an encyclopedia that explains the herbs is displayed. Further, information (location such as URL) for transition to the corresponding web page is associated with the image of the area AR1 and the information of the area AR2.

  For example, when the query “entrance ceremony” is input, information (location and image) related to a web page that sells clothes that can be worn at the entrance ceremony that is the type of service “shopping” is displayed in the area AR1. And the like are displayed on the display unit of the terminal device 10. When a building name or place name such as the query “XX building” is input, the area AR1 includes map information indicating the location of the XX building, which is the service type “map”, and the appearance of the XX building. Is displayed on the display unit of the terminal device 10.

[Process to generate learned model]
Hereinafter, a process for generating the learned model 132 that can determine the type of service corresponding to the input query with higher accuracy will be described.

(Counting clicks)
The evaluation unit 104 generates analysis information 124 based on the action history information 122. FIG. 4 is a diagram illustrating an example of the contents of the analysis information 124. The vertical axis in FIG. 4 indicates the number of clicks, and the horizontal axis in FIG. 4 indicates the year and month. The evaluation unit 104 refers to the action history information 122 and derives the number of times the area AR1 illustrated in FIG. 3 has been clicked for each combination of the query and the service type. For example, the evaluation unit 104 derives the number of times the area AR1 is clicked in the first period P1 (for example, January) and the number of times the area AR1 is clicked in the second period P2 (for example, 1 to December). The second period P2 is a period longer than the first period P1.

(First score)
The evaluation unit 104 derives a first score based on the analysis information 124. For example, the first score is an index indicating the importance of the query to be evaluated in the first period P1 when the second period P2 is targeted. The first score Score1 is derived based on, for example, the following formula (1). “P1_CLICK” is the number of clicks in the first period P1, and “P2_CLICK” is the number of clicks in the second period P2. “P1_CLICK” is an example of a first index, and “P2_CLICK” is an example of a second index.
Score1 = (P1_CLICK / P2_CLICK) (1)

  The following first index may be used instead of (or in addition to) “P1_CLICK”. The first index is, for example, the number of selections that the user has selected a service related to the type of service, the number of conversions that is the number of times the user has performed a predetermined action after the user has selected the service, or the user has selected the service This is an index based on one of the selected unit prices, which is the amount of billing, or a combination thereof.

  Further, instead of (or in addition to) “P2_CLICK”, the following second index may be used. The second index is, for example, the number of selections by which the user has selected a service related to the type of service, the number of conversions that the user has performed a predetermined action after selecting the service, or after the user has selected the service This is an index based on one of the selected unit prices, which is the amount of billing, or a combination thereof.

(Second score)
The evaluation unit 104 derives a second score based on the analysis information 124. For example, when the second period P2 is the target, the second score is an index indicating the importance of the query to be evaluated in the first period P1 in consideration of the first score. The second score Score2 is derived based on the following formula (2), for example. “LG” is, for example, the logarithm of “P2_CLICK”.
Score2 = Score1 × LG (2)

  For example, “LG” is a logarithm with base 10. By multiplying “LG”, a more appropriate score can be derived. This is because by adding “LG”, a high score is derived for a query with a low search frequency or a query with a low number of searches, and is prevented from being determined as a periodic query. Also, by adding “LG”, it is possible to suppress the possibility that a query (such as spam) that the user has not input is determined as a periodic query.

  Instead of (or in addition to) “LG”, the following third index may be used. The third index occurs when the user selects a service related to the type of service, the number of conversions that the user performs a predetermined action after the user selects the service, or after the user selects the service It is a logarithm of an index based on any one of the selected unit prices that are billing amounts or a combination thereof.

  FIG. 5 is a diagram illustrating an example of the content of the score information 126. The above-described processing is performed, a score is derived for a plurality of combinations of queries and service types, and score information 126 is generated. The score information 126 is information in which the query to be evaluated, the first period P1, the second period P2, the first score, and the second score are associated with each other.

(Determining periodic query)
The determination unit 106 determines a periodicity query based on the score information 126. For example, when the second score is equal to or greater than the threshold and the second score is used as a reference, the determination unit 106 determines the upper predetermined number of queries in the ranking as the periodic query.

  Note that the determination unit 106 determines whether the second score (and / or the first score) is equal to or higher than the threshold value, or the top predetermined number of queries in the ranking when the second score (and / or the first score) is used as a reference. May be determined as a periodic query. In addition, the determination unit 106 statistically processes one or more items among the first score, the second score, and the ranking result based on the first score or the second score to perform a periodic query. May be determined.

  In addition, the determination unit 106 may determine a periodic query based on the third score. The third score is a score derived based on the first index and the second index, for example. Further, for example, the evaluation unit 104 inputs one or more of the first to third indices into a predetermined function, and the predetermined function evaluates the periodic query based on the output result. Good.

[Flowchart (Part 1)]
FIG. 6 is a flowchart illustrating an example of a flow of processing executed by the information processing apparatus 100. This process is a process when a periodic query is determined. First, the evaluation unit 104 acquires the action history information 122 (S10), and analyzes the acquired action history information 122 (S12). Thereby, the analysis information 124 is generated.

  Next, the evaluation unit 104 derives a first score based on the analyzed analysis information 124 (S14). Next, the evaluation unit 104 derives a second score based on the analysis information 124 and the first score (S16).

  Based on the derived second score, the determination unit 106 derives a top predetermined number of queries having a second score equal to or higher than the threshold as a periodic query, and the derived periodic query is represented by the periodic query information. 128 is stored in the storage unit 120 (S18). Thereby, the process of one routine of this flowchart is completed.

  Through the processing described above, the information processing apparatus 100 can derive a periodic query with high accuracy.

  FIG. 7 is a diagram illustrating an example of a periodic query derived by the information processing apparatus 100. For example, the information processing apparatus 100 derives a monthly periodic query in the past several years. In addition, for example, the determination unit 106 has a second score that is equal to or higher than a threshold value in the past several years in a predetermined period (for example, a predetermined month such as January, a predetermined season, etc.), and cycles the upper predetermined number of queries. It may be extracted as a sex query. For example, in January in the past three years, when the query “Nanagusa” has a second score equal to or higher than the threshold and the upper predetermined number of queries are periodic queries, the query may be determined as a periodic query. This excludes queries whose number of inputs or searches has temporarily increased due to a trend or the like.

[Flowchart (Part 2)]
FIG. 8 is a flowchart illustrating another example of the flow of processing executed by the information processing apparatus 100. This process is a process when the learned model 132 is generated. In the following description, the first period or the second period is set based on the month or year, but the present invention is not limited to this. For example, the first period is a predetermined period in which an event occurs (a period in which a product is released or an Olympic period), and the second period is a predetermined period before (or after) the event occurs. It may be.

  First, the learning unit 108 acquires learning data 130 for a predetermined period (for example, two months) before the target month (S100). Next, the learning unit 108 acquires learning data 130 related to the periodic query (S102). Next, the learning unit 108 performs learning based on the learning data 130 acquired in S100 and S102 (S104). FIG. 9 is a diagram illustrating an example of the content of the learning data 130. The learning data 130 for two months before the target month is information in which a query input during that period is associated with a service type. The learning data 130 related to the periodic query is information in which a query input in the past January (first period) is associated with a service type.

  Returning to FIG. Next, the learning unit 108 generates a learned model 132 (S106). The learned model 132 is generated using, for example, a neural network, SVM, or the like as a constituent element. Thereby, the process of one routine of this flowchart is completed.

  Through the above-described processing, a learned model 132 as shown in FIG. 10 is generated. FIG. 10 is a diagram illustrating an example of the learned model 132. For example, the learned model 132 is a model that determines whether or not a given service type corresponds to a query input. The learned model 132 is generated for each type of service to be determined, for example. The learned model 132 is a model that determines, for example, whether the service type is a recipe, the service type is shopping, or the service type is a map. For example, when the query “Nanakusa” is input to the learned model 132 for determining whether the service type is a recipe, the learned model 132 determines that the query corresponds to the recipe service type, and the determination Output the result.

[Flowchart (Part 3)]
FIG. 11 is a flowchart illustrating another example of the flow of processing executed by the service server 20 and the information processing apparatus 100. This process is an example of a process flow in which when the user inputs a query, the service type specific information corresponding to the query is provided to the user.

  First, the service server 20 determines whether a query has been input (S200). Then, the service server 20 transmits the input query to the information processing apparatus 100.

  Next, the processing unit 110 of the information processing apparatus 100 uses the learned model 132 to determine the type of service corresponding to the transmitted query (S202). The processing unit 110 transmits the determined service type to the service server 20. Next, the search unit 22 of the service server 20 searches the network NW for information related to the query (S204). Next, the service server 20 determines a link to be included in the search result display screen based on the service type determined in S202, and associates the determined link with the shortcut area (S206). In this way, the service server 20 generates information related to the search result based on the service type. The information related to the search result is, for example, a link to be included in the search result display screen, such as a link to an image, a link to a text, or a combination of a link to an image and a link to a text. And the service server 20 transmits the produced | generated information to the terminal device 10, and the image (specific information) based on the produced | generated information is displayed on the display part of the terminal device 10. FIG. Thereby, the process of one routine of this flowchart is completed.

  Through the processing described above, the information processing apparatus 100 can determine a more appropriate service type. Then, the service server 20 can perform marketing more appropriately based on the determined service type, or can provide useful information for the user.

[Comparative example]
For example, the learned model 132 is generated for each target month (or period). For example, the learned model used in January may be generated using information from the most recent months (for example, November and December) in order to suppress the learning cost. For example, learning data is generated based on queries input in November and December, information displayed in the area AR1 according to the query, frequency of clicking information displayed in the area AR1, and the like. The learning data is information in which a query and a service type corresponding to the query are associated with each other.

  A learned model is generated using this learning data, but this learned model does not take into account periodic queries that are frequently input in the target month. For this reason, for example, in January, when the query “XXX” is input, the service type is determined as “shopping”, and it is more likely that the user will be provided with information related to shopping. In addition, the service type is determined as “recipe” or “map”, and information regarding these service types is provided to the user, or appropriate information is provided to the user without determining the service type. It may not be possible.

  Specifically, in January, when the query “Nanagusa” is input, the service type is determined as “recipe”, and the information about the recipe is provided to the user, and the probability of being clicked is higher. However, as shown in the image IM1 in FIG. 12, the service type is determined as “shopping” or “map”, or as shown in the image IM2 in FIG. 12, the service type is not determined in the area AR1. There are cases where information according to the type of service cannot be provided.

  In this way, when only the information of the past few months is used as learning data, for example, the number of times and frequency of inputting a query that becomes a periodic query is low. The content of the type is not reflected, and marketing may not be performed more appropriately, or information useful for the user may not be provided.

  On the other hand, in the present embodiment, by determining a periodic query and including the periodic query and the type of service of the periodic query in the learning data, it is possible to perform marketing more appropriately or be beneficial to the user. A learned model 132 can be generated that can provide information. Specifically, in order to generate the learned model 132 used in January, learning data including a periodic query determined based on information in the past January and a service type corresponding to the periodic query are included. This is because it is used.

  According to the first embodiment described above, the information processing apparatus 100 is an action of a user with respect to a service provided in response to a query input by the user in the first period and the second period. A query that a user focuses on periodically can be accurately extracted by evaluating a query based on a user's behavior to be managed and determining a periodic query having a predetermined periodicity based on an evaluation result. it can.

Second Embodiment
Hereinafter, a second embodiment will be described. In the first embodiment, information on the type of service determined based on the query is displayed in the shortcut area, and the score is derived based on the user's action on this information. On the other hand, in the second embodiment, the shortcut area is not provided, and the user's behavior for the service provided in response to the query is based on the user's behavior managed for each service type. A score is derived. Hereinafter, a second embodiment will be described.

  The service server 20 provides information related to the query input by the user to the terminal device 10. FIG. 13 is a diagram illustrating an example of an image IM3 displayed on the display unit of the terminal device 10 by the service server 20 in the second embodiment. If the query is “Nanagusa”, the image IM3 is provided with, for example, a link L1 of a web page introducing a recipe for herbs, a link L2 of a web page selling a product of “Nanagusa”, and a herb. The link destination L3 etc. of the store which is doing is included. The storage unit 120 stores correspondence information. The correspondence information is information in which the link destinations (L1 to L3), the area where the link destination is displayed, and the service type are associated with each other. The service type is, for example, a service type determined based on link destination HTML (Hyper Text Markup Language), character information included in the link destination web page, or the like.

  Based on the correspondence information, the processing unit 110 acquires an area (link destination) clicked by the user and the type of service at the link destination, and stores the acquired information in the storage unit 120 as action history information 122. Then, the evaluation unit 104 generates analysis information 124 </ b> A based on the action history information 122. FIG. 14 is a diagram illustrating an example of the analysis information 124A of the second embodiment. The analysis information 124A is, for example, information in which a target period for which an action history is acquired, a service type, and the number of clicks are associated with a query. The evaluation unit 104 derives the first score and the second score by a process equivalent to the process of the first embodiment.

  According to the second embodiment described above, the information processing apparatus 100 manages user behavior for a search result based on a query (user behavior for a service provided in response to a query, which is managed for each type of service. By evaluating the query based on the user's behavior), the same effect as that of the first embodiment can be obtained.

<Third Embodiment>
Hereinafter, the third embodiment will be described. In the first embodiment, the query is evaluated based on the user's behavior with respect to the information of the first service type associated with the query input by the user in the first period and the second period. On the other hand, in the third embodiment, the information processing apparatus 100 refers to correspondence information in which a service type is associated with a query, and is a query input by a user in the first period. Based on the number of queries input associated with the service type in the query and the number of queries entered by the user in the second period and associated with the service type in the correspondence information. To evaluate. Hereinafter, the third embodiment will be described.

  In the third embodiment, correspondence information in which a query is associated with a service type is stored in the storage unit 120. Then, the evaluation unit 104 generates analysis information 124B based on the correspondence information and the action history information 122. FIG. 15 is a diagram illustrating an example of the content of the analysis information 124B of the third embodiment. The vertical axis in FIG. 15 indicates the number of searches, and the horizontal axis in FIG. 15 indicates the year and month. The evaluation unit 104 refers to the analysis information 124B, and derives the number of times a predetermined query (for example, “Nanagusa”) has been searched. For example, the evaluation unit 104 derives the number of times that a predetermined query has been searched in the first period P1 (for example, January) and the number of times that the predetermined query has been searched for in the second period P2 (for example, 1 to December). .

(First score)
The evaluation unit 104 derives a first score based on the analysis information 124B. The first score Score1 is derived based on the following formula (3), for example. “P1_SEARCH” is the number of searches in the first period P1, and “P2_SEARCH” is the number of searches in the second period P2. “P1_SEARCH” is another example of the first index, and “P2_SEARCH” is another example of the second index.
Score1 = (P1_SEARCH / P2_SEARCH) (3)

(Second score)
The evaluation unit 104 derives a second score based on the analysis information 124B. The second score Score2 is derived based on, for example, the following formula (4). “LG” is the logarithm of “P2_SEARCH”. “P2_SEARCH” is another example of the third index.
Score2 = Score1 × LG (4)

  As described above, the evaluation unit 104 generates score information 126 and determines a periodicity query based on the score information 126. Then, the determination unit 106 determines a periodic query based on the score information 126.

  According to the third embodiment described above, the information processing apparatus 100 is a query input by the user in the first period and the second period, and the service information in the correspondence information in which the service type is associated with the query. The same effect as the effect of the first embodiment is obtained by evaluating the query based on the number of inputs of the query associated with the type and determining the periodic query having a predetermined periodicity based on the evaluation result. Can be played.

  According to the embodiment described above, the information processing apparatus 100 is a user action on a service provided in response to a query input by the user in the first period, and is managed for each type of service. And an evaluation unit 104 that evaluates a query based on a user's behavior with respect to a service provided in response to a query input by the user in the second period and managed for each type of service; By including the determination unit 106 that determines a periodic query having a predetermined periodicity based on the evaluation result of the evaluation unit 104, it is possible to accurately extract a query that the user pays attention to periodically.

  According to another embodiment described above, the information processing apparatus 100 refers to correspondence information in which a service type is associated with a query, and is a query input by a user in a first period, The number of queries in which the service type is associated in the correspondence information in which the service type is associated with the query and the query input by the user in the second period, and the service type corresponds in the correspondence information By providing an evaluation unit 104 that evaluates a query based on the number of attached query inputs, and a determination unit 106 that determines a periodic query having a predetermined periodicity based on the evaluation result of the evaluation unit 104 Queries that the user focuses on periodically can be extracted with high accuracy.

  As mentioned above, although the form for implementing this invention was demonstrated using embodiment, this invention is not limited to such embodiment at all, In the range which does not deviate from the summary of this invention, various deformation | transformation and substitution Can be added.

DESCRIPTION OF SYMBOLS 1 ... Information processing system, 10 ... Terminal device, 20 ... Service server, 100 ... Information processing device, 102 ... Communication part, 104 ... Evaluation part, 106 ... Determination part, 108 ... Learning part, 110 ... Processing part 110, 120 ... Storage unit 122... Action history information 124. Analysis information 126 126 Score information 128 Periodic query information 130 Learning data 132 Learning model

Claims (12)

User behavior for a service provided in response to a query input by the user in the first period and managed for each type of service, and response to the query input by the user in the second period An evaluation unit that evaluates the query based on a user's behavior with respect to the service provided and managed by each type of the service, and
A determination unit that determines a periodic query having a predetermined periodicity based on an evaluation result of the evaluation unit;
An information processing apparatus comprising: The second period is longer than the first period.
The information processing apparatus according to claim 1. The evaluation unit evaluates the query based on a first index obtained by indexing the user's behavior in the first period;
The first index is the number of selections by which the user has selected a service related to the type of service, the number of conversions by which the user has performed a predetermined action after the user has selected the service, or the user has selected the service A metric based on one or more of the selected bids that are accrued after
The information processing apparatus according to claim 1 or 2. The evaluation unit evaluates the query based on a second index obtained by indexing the user's behavior in the second period;
The second index is the number of selections that the user has selected a service related to the type of service, the number of conversions that is the number of times the user has performed a predetermined action after the user has selected the service, or the user has selected the service A metric based on one or more of the selected bids that are accrued after
The information processing apparatus according to any one of claims 1 to 3. The evaluation unit evaluates the query based on a third index that indexes the user's behavior in the second period,
The third index is the number of selections that the user has selected a service related to the type of service, the number of conversions that is the number of times the user has performed a predetermined action after the user has selected the service, or the user has selected the service Is the logarithm of the metric based on one of the selected bids, or any combination of these,
The information processing apparatus according to any one of claims 1 to 4. Using the data in which the periodic query and the service type are associated with each other as learning data, a learned model is generated that is used to determine which service type information should be provided to the user when a query is input. A learning unit;
The information processing apparatus according to any one of claims 1 to 5. The learning data further includes data in which a query input by a user in the most recent period when the learned model is generated and the service type are associated with each other.
The information processing apparatus according to claim 6. A processing unit that inputs a query input by a user to the learned model and determines a type of service to be provided to the user based on an output result output by the learned model;
The information processing apparatus according to claim 6 or 7. The information processing apparatus according to any one of claims 6 to 8,
A query input by a user is transmitted to the information processing device, and the information processing device inputs the query to the learned model and obtains a service type determined based on an output result output from the learned model. A service server that provides information about search results based on the type of service acquired to the user;
An information processing system comprising: Refer to the correspondence information in which the service type is associated with the query,
The number of queries input by the user in the first period and associated with the service type in the correspondence information;
An evaluation unit that evaluates the query based on the number of queries input by the user in the second period and associated with the service type in the correspondence information;
A determination unit that determines a periodic query having a predetermined periodicity based on an evaluation result of the evaluation unit;
An information processing apparatus comprising: Computer
User behavior for a service provided in response to a query input by the user in the first period and managed for each type of service, and response to the query input by the user in the second period The user's behavior with respect to the service provided and the user behavior managed for each type of the service, and the query is evaluated,
Determining a periodic query having a predetermined periodicity based on the evaluation result;
Information processing method. On the computer,
User behavior for a service provided in response to a query input by the user in the first period and managed for each type of service, and response to the query input by the user in the second period The user's behavior with respect to the service provided and the user behavior managed for each type of the service, the query is evaluated,
A periodic query having a predetermined periodicity is determined based on the evaluation result;
program.

Images