JP2019053643A - Information processing device and program - Google Patents

Abstract

To match a user's attribute with a recommendation object and with context information about a context in using the recommendation object in the case of recommending an object to the user.SOLUTION: A presentation part 105 selects a context having a correlation with a user's attribute among contexts having a correlation with each item, and presents a recommendation item being an item having a correlation with the selected context and information about a context in using the recommendation item. Also, the presentation part 105 selects an item having a correlation with the user's attribute among items having a correlation with each context, and presents a recommendation item being the selected item and information about a context in using the recommendation item.SELECTED DRAWING: Figure 2

Description

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

  Regarding a technique for recommending products and services (hereinafter collectively referred to as items) to a user, for example, Patent Document 1 discloses that an item is recommended according to the environment. Patent Document 2 discloses that an item is recommended in accordance with the situation when the situation changes. As described above, an item recommendation mechanism according to the environment and situation in which the item is used (these are also referred to as contexts) is already known.

Special table 2004-536488 gazette Japanese Patent No. 5979159

  By the way, even if there is a context suitable for the user, there may be no item suitable for the context, or there may be no context suitable for the item even if there is an item suitable for the user.

  Therefore, the present invention has an object to match a user attribute, a recommended object, and situation information about a situation when using the recommended object when recommending the object to the user.

  The invention according to claim 1 is a first analysis unit that analyzes a relationship between a user attribute and an object attribute; and a relationship between the attribute of the object and a situation when the object is used. A second analysis unit for analyzing; a third analysis unit for analyzing relevance between the user attribute and the situation when the object is used; the first analysis unit; the second analysis unit; According to the analysis result by the three analysis units, according to the situation information regarding the recommended object that is an object recommended for the user and the situation when the recommended object is used, each of the relevance is Provided is an information processing apparatus comprising a presentation unit that presents the highest recommended object and situation information.

  According to a second aspect of the present invention, in the configuration according to the first aspect, the presenting unit selects the situation relevant to the attribute of the user from the situations relevant to each of the objects. The recommended object that is the object related to the selected situation and the situation information regarding the situation when the recommended object is used are presented.

  According to a third aspect of the present invention, in the configuration according to the first aspect, the presenting unit selects the target object related to the user attribute from the target objects related to the respective situations. Then, the recommended object that is the selected object and the situation information regarding the situation when the recommended object is used are presented.

  According to a fourth aspect of the present invention, in the configuration according to the first aspect, the presentation unit selects the set of the object and the situation that are related to each other, and among the selected set of the object and the situation The target object and the situation set relevant to the attribute of the user are identified, and the recommended object that is the object and the situation information regarding the situation when the recommended object is used are presented.

  According to a fifth aspect of the present invention, in the configuration according to the first aspect, the presenting unit selects the target object related to the user attribute and the situation related to the user attribute, respectively. Then, among the selected object and the situation, the object and the situation that are related to each other are specified, and the recommended object that is the object and the situation regarding the situation when the recommended object is used Present information.

  The invention according to claim 6 is the configuration according to any one of claims 1 to 5, wherein the presentation unit includes the attribute of the user and the object, the situation and the attribute of the user, and the situation and the Weighting is performed on the relevance of each object, and the recommended information to be presented and the situation information regarding the situation when the recommended object is used are selected.

  According to a seventh aspect of the present invention, in the configuration of the sixth aspect, the presenting unit learns the weighting parameter in accordance with a user designation.

  According to an eighth aspect of the present invention, there is provided a computer comprising: a first analysis unit that analyzes a relationship between a user attribute and an object attribute; and an attribute of the object and a situation when the object is used A second analysis unit for analyzing relevance, a third analysis unit for analyzing relevance between the attribute of the user and the situation when the object is used, the first analysis unit, and the second analysis unit And according to the analysis result by the third analysis unit, according to, the recommended object which is the object recommended for the user and the situation information regarding the situation when the recommended object is used, Provided is a program for functioning as a presenting unit for presenting recommended objects and situation information having the highest relevance.

According to the first and eighth aspects of the invention, when recommending an object to a user, the user's attributes, the recommended object, and the situation information regarding the situation when using the recommended object can be matched. .
According to the invention which concerns on Claim 2, it selects from the situation according to a user.
According to the invention which concerns on Claim 3, it selects from the target object according to the user.
According to the invention which concerns on Claim 4, it selects from the group of the target object and the situation which have relevance.
According to the invention which concerns on Claim 5, it selects from the object according to a user, and the condition according to a user.
According to the invention of claim 6, recommendation according to weighting is realized.
According to the invention of claim 7, appropriate weighting is learned.

2 is a diagram illustrating a hardware configuration of the information processing apparatus 1. FIG. 2 is a block diagram showing a functional configuration of the information processing apparatus 1. FIG. 5 is a flowchart showing the operation of the information processing apparatus 1. 5 is a flowchart showing the operation of the information processing apparatus 1. 5 is a flowchart showing the operation of the information processing apparatus 1. 5 is a flowchart showing the operation of the information processing apparatus 1. The figure which illustrated the relationship between the attribute of a user, the attribute of an item thing, and context.

An example of an embodiment of the present invention will be described.
<Configuration>
FIG. 1 is a diagram illustrating a hardware configuration of the information processing apparatus 1. The information processing apparatus 1 is an example of an information processing apparatus according to the present invention. The information processing apparatus 1 is a computer including a control unit 11, a storage unit 12, a communication unit 13, and a UI (User Interface) unit 14. The control unit 11 includes an arithmetic device such as a CPU (Central Processing Unit) and a storage device such as a ROM (Read Only Memory) and a RAM (Random Access Memory), and performs overall control of the information processing device 1. Do. The ROM stores firmware that describes the startup procedure of hardware and an OS (Operating System). The RAM is used for storing data when the CPU executes an operation.

  The storage unit 12 includes, for example, a semiconductor memory or a hard disk storage device, and stores an OS and programs for realizing various functions. Furthermore, the storage unit 12 stores information related to a result of using a shopping website on the Internet using a communication device such as a personal computer or a smart phone by a plurality of users.

  Specifically, the storage unit 12 stores a purchase history (for example, a user of a user who purchased an item) that is a history of each user purchasing a product or service (hereinafter collectively referred to as an item) on a shopping website on the Internet. ID, user attribute, purchase date and time, purchased item name, item purchase situation, and item attribute such as category, price, or usage)). The purchase history may be collected from the web server device that manages the shopping website on the Internet to the information processing device 1 via the Internet, or the information processing device 1 itself manages the shopping website on the Internet. It may be a web server device.

  The communication unit 13 includes a communication I / F (Interface) for performing communication with an external communication device, and is connected to a communication line such as the Internet. The UI unit 14 includes a display unit and an operation unit. The display unit includes, for example, a liquid crystal display device, and displays a screen for operating the information processing apparatus 1 on the display surface. The operation unit includes, for example, a touch panel provided so as to cover the display surface of the display unit, a keyboard provided at a position adjacent to the display surface, etc., and receives an operation and controls a signal according to the operation. To the unit 11. The control unit 11 controls the information processing apparatus 1 according to the contents of this operation.

  FIG. 2 is a block diagram illustrating a functional configuration of the information processing apparatus 1. These functions are realized when the control unit 11 executes a program stored in the storage unit 12. The acquisition unit 101 acquires a user attribute, an item attribute, and a situation when the item is used (for example, including a use purpose and a use environment, hereinafter referred to as a context). The first analysis unit 102 analyzes the relationship between the user attribute and the item attribute. The second analysis unit 103 analyzes the relationship between the attribute of the item and the context when the item is used. The third analysis unit 104 analyzes the relationship between the user attribute and the context when the item is used. The presenting unit 105 uses a recommended item that is an item recommended for the user and the recommended item when the first analysis unit 102, the second analysis unit 103, and the third analysis unit 104 are used. Information related to the context (situation information) is transmitted to the user communication device. More specifically, the presentation unit 105 is a situation information regarding a recommendation object that is an object recommended to the user and a situation when the recommendation object is used in each analysis result described above. Each of the recommended objects and the situation information having the highest relevance is transmitted to the user's communication device, and these are presented to the user.

<Operation>
Next, the operation of the information processing apparatus 1 will be described with reference to FIGS.
<Operation example 1>
First, an example in which a context is first selected in order to select an item to be recommended will be described. Each of the first analysis unit 102, the second analysis unit 103, and the third analysis unit 104 uses the information acquired by the acquisition unit 101 (here, the purchase history stored in the storage unit 12), and the user attribute and the item. The relationship between attributes, the relationship between item attributes and context, and the relationship between user attributes and context are analyzed (step S101). For example, when the 1st analysis part 12 calculates | requires the relationship of a user attribute and an item attribute,
Number of simultaneous occurrences of user attribute X and item attribute Y in purchase history / number of occurrences of either = | X∩Y | / | X∪Y |
The relationship between the two is obtained by the mathematical formula. In addition to this, for example, paying attention to the mutual information amount of the user attribute X and the item attribute Y, the mathematical expression P (X, Y) log (P (X, Y) / P (X) P (Y)) is used. The relationship between the two may be obtained. Further, the relevance between the two may be obtained by an inner product of the vectors when projected onto the latent vector by a method such as matrix decomposition or tensor decomposition. Further, the relationship between the item attribute and the context and the relationship between the user attribute and the context can be obtained by the same method.

  Here, FIG. 7 illustrates an example of the above analysis result for the user attribute of a male college student without her. In the example of FIG. 7, the relationship between this user attribute and the item Italian restaurant A is 0.4, and the relationship between this user attribute and the item Chinese restaurant B is 0.3, The relationship between this user attribute and the item Japanese restaurant C is 0.3, and the relationship between this user attribute and the item Yakiniku restaurant D is 0.5. This user attribute and the hamburger store E The relevance of this item is 0.2, and the relevance of this user attribute to the item okonomiyaki F is 0.7. The relevance takes a value of 0 or more and 1 or less, and the larger the value, the higher the relevance.

  In FIG. 7, the relationship between this user attribute and the context of “one person” is 0.5, and the relationship between this user attribute and the context of “date” is 0.2. The association with the context of the party is 0.0, the relation between this user attribute and the context of the joint party is 0.7, and the relation between this user attribute and the context of the banquet is 0.3. The relevance between this user attribute and the context of a lunch meeting is 0.1.

  In FIG. 7, for example, the relevance between the item Italian restaurant A and the context of single meal is 0.2, the relevance of the item Chinese restaurant B and the context of date is 0.5, The relationship between the item Japanese restaurant C and the context of girls' association is 0.3, the relationship between the item Yakiniku restaurant D and the context of joint party is 0.05, the item hamburger store E and the banquet The relationship with the context is 0.05, and the relationship between the item okonomiyaki F and the context of the lunch meeting is 0.05.

  Returning to FIG. 3, the presentation unit 105 selects a context whose relevance to any item is equal to or greater than a threshold (step S <b> 102). In the example of FIG. 7, when the relevance threshold is 0.3, contexts such as a meal, a date, a girls' association, and a lunch meeting are selected.

  Next, the presentation unit 105 selects a context whose relevance to the user attribute is equal to or greater than a threshold among the selected contexts (step S103). In the example of FIG. 7, if the relevance threshold is 0.5, the context of “one person” is selected.

  Next, the presentation unit 105 selects an item having the highest relevance with the selected context (step S104). Thereby, in the example of FIG. 7, an item called Yakiniku restaurant D is selected. Then, the presentation unit 105 presents the selected item and information regarding the selected context to the user. In the example of FIG. 7, for example, the user is notified of a message “I recommend you a Yakiniku restaurant D for one person”.

  In this way, the presentation unit 105 selects a context that is related to the user attribute from the contexts that are related to the attribute of each item, and the recommended item that is an item that is related to the selected context and the relevant item Information about the context when the recommended item is used is presented. Note that the items that are candidates for selection are not all items, but items that are profitable for businesses that use the items (for example, there are advertisers), the current location of the user, or the user's location. The frequency may be narrowed down to a value close to a threshold or more. Moreover, you may select with the weighted sum of the relationship with a context, and the relationship with a user attribute.

<Operation example 2>
Next, an example in which an item is first selected in order to select an item to be recommended will be described. The first analysis unit 102, the second analysis unit 103, and the third analysis unit 104 respectively determine the relationship between the user attribute and the item attribute, the relationship between the item attribute and the context, and the relationship between the user attribute and the context. Analysis is performed (step S201, the same processing as step S101 described above).

  The presenting unit 105 selects items whose relevance with any context is greater than or equal to a threshold (step S202). In the example of FIG. 7, if the relevance threshold is 0.3, the items Italian restaurant A, Chinese restaurant B, Japanese restaurant C, and yakiniku restaurant D are selected.

  Next, the presentation unit 105 selects a context whose relevance to the user attribute is equal to or greater than a threshold among the selected items (step S203). In the example of FIG. 7, if the relevance threshold is 0.5, the item “Yakiniku restaurant D” is selected.

  Next, the presentation unit 105 selects a context having the highest relevance with the selected item (step S204). Thereby, in the example of FIG. 7, the context of “one person” is selected. Then, the presentation unit 105 presents the selected item and information regarding the selected context to the user. In the example of FIG. 7, the user is notified of a message “I recommend you a Yakiniku restaurant D for one person”.

  As described above, the presentation unit 105 selects an item that is related to the user attribute from items related to each context, and the recommended item that is the selected item and the recommended item are used. Present information about the context. Note that the items that are candidates for selection are not all items, but items that are profitable for businesses that use the items (for example, there are advertisers), the current location of the user, or the user's location. The frequency may be narrowed down to a value close to a threshold or more. Moreover, you may select with the weighted sum of the relationship with a context, and the relationship with a user attribute.

<Operation example 3>
Next, an example in which an item and context pair is first selected to select an item to be recommended will be described. The first analysis unit 102, the second analysis unit 103, and the third analysis unit 104 respectively determine the relationship between the user attribute and the item attribute, the relationship between the item attribute and the context, and the relationship between the user attribute and the context. Analysis is performed (step S301, the same processing as step S101 described above).

  The presenting unit 105 selects a pair in which the relationship between the item and the context is greater than or equal to a threshold (step S302). In the example of FIG. 7, when the relevance threshold is 0.5, a set of Italian restaurant A and girls' association, Chinese restaurant B and date, yakiniku restaurant D, and single meal is selected.

  Next, the presenting unit 105 selects, from among the selected pairs, the pair having the highest relevance among the pairs in which the relevance between the user attribute and any of the items and contexts constituting the pair is the threshold value or more. (Step S303). In the example of FIG. 7, if the relevance threshold is 0.5, a set of yakiniku restaurant D and single meal is selected. Then, the presentation unit 105 presents the items that make up the selected set and information about the context to the user. In the example of FIG. 7, the user is notified of a message “I recommend you a Yakiniku restaurant D for one person”.

  In this way, the presentation unit 105 selects a pair of items and contexts that are related to each other, identifies a pair of items and contexts that are related to the user attribute from the selected sets of items and contexts, and Information regarding a recommended item that is an item and context when the recommended item is used is presented.

<Operation example 4>
Next, an example will be described in which an item associated with a user attribute and a context associated with the user attribute are first selected to select an item to be recommended. The first analysis unit 102, the second analysis unit 103, and the third analysis unit 104 respectively determine the relationship between the user attribute and the item attribute, the relationship between the item attribute and the context, and the relationship between the user attribute and the context. Analysis is performed (step S401, the same processing as step S101 described above).

  The presenting unit 105 selects an item whose relevance to the user attribute is equal to or higher than the threshold, and also selects a context whose relevance to the user attribute is equal to or higher than the threshold (step S402). In the example of FIG. 7, if the relevance threshold is 0.5, the yakiniku restaurant D is selected as an item whose relevance to the user attribute is equal to or higher than the threshold, and the lone meal is used as the context whose relevance to the user attribute is equal to or higher than the threshold. And Gokon are selected.

  Next, the presenting unit 105 selects a pair having the highest relevance among the items and contexts whose mutual relevance is equal to or greater than the threshold among the selected items and contexts (step S403). In the example of FIG. 7, if the relevance threshold is 0.5, a set of yakiniku restaurant D and single meal is selected. Then, the presenting unit 105 presents the selected item and information regarding the context to the user. In the example of FIG. 7, the user is notified of a message “I recommend you a Yakiniku restaurant D for one person”.

  In this way, the presentation unit 105 selects an item related to the user attribute and a context related to the user attribute, and selects an item and context related to each other from the selected item and context. It identifies and presents the recommended item that is the item and information about the context when the recommended item is used.

  According to the present embodiment, when an item is recommended to the user, the recommendation can be performed in a state where the user attribute, the item attribute, and the information regarding the situation when the item is used are matched.

<Modification>
The embodiment may be modified as follows. A plurality of modified examples may be combined.
<1>
The presentation unit 105 weights the user attribute and item attribute, the context and the user attribute, and the relevance of the context and the item attribute, and relates to the context when the recommended item to be presented and the recommendation target are used. Information may be selected. Furthermore, the presentation unit 105 may learn a weighting parameter in accordance with a user designation.

For example, as an example of weighting,
It is conceivable that the weight value for the relation between the user attribute and the context> the weight value for the relation between the context and the item attribute ≧ the weight value for the relation between the user attribute and the item attribute. In the first place, the recommendation that puts weight on the context when using an item is effective. This weighting may be machine-learned from an action history (for example, a purchase history after recommendation) when a user responds to an item recommendation. In addition, since each weight is considered to be different for each user or for each item, machine learning may be performed so that the weight is different for each user or item. This is because, for example, there are users who place importance on the consistency between the context and the item, and there are users who place importance on whether or not the context is the item that suits them.

<2>
In addition to the method described in the embodiment, a method for selecting a context or an item may be based on the following evaluation score.
Evaluation score of context A = Relevance between user attribute p and context A + max / average (relationship between context A and each item)
Item A evaluation score = Relevance between user attribute p and item A + max / average (relationship between item A and each context)

<3>
In the embodiment, the user's preference is learned based on the result of the user using a shopping website on the Internet, but the user's preference is determined based on the result of using a store that sells real items. You may learn. Specifically, the storage unit 12 is a purchase history (for example, a user ID of a user who purchased a product, a purchase date and time, a name of the purchased product, a price, a usage, etc.) Product attributes) and browsing history (user ID, browsing date and identification information of browsed product shelves), which is a history of browsing products by each user in the store, are stored. The purchase history is managed using, for example, a store card given to each user. The browsing history is managed, for example, by analyzing an image captured by a shopping cart or a camera mounted in a store.

<4>
In the present invention, the determination of whether or not there is a relationship may be, for example, a determination as to whether or not the relationship is greater than or equal to a threshold value, or may be a determination as to whether or not the relationship is maximum. For example, in the first operation example, the presentation unit 105 selects a context having the maximum relevance to the user attribute from the selected contexts after selecting the context having the relevance to any item or more (step S102). You may make it do (step S103).

<5>
In the above embodiment, the example in which the control unit 11 realizes the above function by executing a program has been described. However, the above function may be implemented by a hardware circuit. The program may be provided by being recorded on a computer-readable recording medium such as an optical recording medium or a semiconductor memory, and the program may be read from the recording medium and installed in the information processing apparatus. In addition, this program may be provided through a telecommunication line.

DESCRIPTION OF SYMBOLS 1 ... Information processing apparatus, 11 ... Control part, 12 ... Memory | storage part, 13 ... Communication part, 14 ... UI part, 101 ... Acquisition part, 102 ... 1st analysis part, 103 ... 2nd analysis part, 104 ... 3rd analysis Part, 105 ... presentation part.

Claims (8)

A first analysis unit that analyzes a relationship between a user attribute and an object attribute;
A second analysis unit for analyzing the relationship between the attribute of the object and the situation when the object is used;
A third analysis unit that analyzes the relevance between the attribute of the user and the situation when the object is used;
In the analysis results by the first analysis unit, the second analysis unit, and the third analysis unit, the recommended object that is a target object recommended for the user and the situation when the recommended target object is used An information processing apparatus comprising: a recommendation unit that presents information and a recommendation object having the highest relevance and status information. The presenting unit
Selecting the situation relevant to the attribute of the user from among the situations relevant to each of the objects;
The information processing apparatus according to claim 1, wherein the recommended object which is the object related to the selected situation and situation information regarding the situation when the recommended object is used are presented. The presenting unit
From among the objects related to each of the situations, select the objects related to the user's attributes;
The information processing apparatus according to claim 1, wherein the recommended object that is the selected object and the situation information related to the situation when the recommended object is used are presented. The presenting unit
Selecting the set of objects and situations that are related to each other;
Among the selected set of the object and the situation, the object and the set of the situation that are related to the attribute of the user are specified, and the recommended object and the recommended object that are the objects are used. The information processing apparatus according to claim 1, wherein situation information relating to a situation at the time of the operation is presented. The presenting unit
Selecting each of the objects related to the user's attributes and the situation related to the user's attributes;
Among the selected object and the situation, the object and the situation that are related to each other are specified, and the recommended object that is the object and the situation information regarding the situation when the recommended object is used The information processing apparatus according to claim 1, wherein the information processing apparatus is presented. The presenting unit
The recommended object to be presented and its recommended object are used by weighting the user attribute and the object, the situation and the user attribute, and the relevance of the situation and the object. The information processing apparatus according to any one of claims 1 to 5, wherein situation information relating to a situation at the time is selected. The presenting unit
The information processing apparatus according to claim 6, wherein the weighting parameter is learned in accordance with a user designation. Computer
A first analysis unit that analyzes a relationship between a user attribute and an object attribute;
A second analysis unit for analyzing the relationship between the attribute of the object and the situation when the object is used;
A third analysis unit that analyzes the relevance between the attribute of the user and the situation when the object is used;
According to the analysis results by the first analysis unit, the second analysis unit, and the third analysis unit, a recommended object that is an object recommended for the user and the recommended object are used. A program for functioning as a presentation unit for presenting the recommended object and the situation information having the highest relevance information.

Images