Information processing device, information processing method, and program

Abstract

This invention provides an information processing device, information processing method, and program that can accurately grasp the deep-seated needs of each user using a new method and provide each user with the most suitable information. [Solution] The information processing device 10 includes a storage unit 12 that stores deep-rooted product needs information associated with each of a plurality of products; a recommendation score calculation unit 111 that calculates a recommendation score for each of the plurality of products based on at least one of the behavioral history information and user attribute information relating to the behavior of users browsing the website; a recommendation product information determination unit 112 that determines the ranking of products to recommend to the user based on the recommendation scores; and a user deep-rooted needs estimation unit 113 that estimates deep-rooted needs information indicating the desires of each user based on at least one of the recommendation score and ranking of each product included in a predetermined number of products that are ranked highly, and the deep-rooted product needs information.

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] Conventionally, there has been known a technique of acquiring information on the search history and site browsing history of site visitors who browse a website, and identifying the fields of interest of the site visitors or displaying advertisements in those fields based on the history information. For example, Patent Document 1 discloses a system that stores and manages the history of articles viewed by a user and related fields, and displays advertisements in the order of priority determined based on the user's field of interest and browsing history.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] According to the technique of Patent Document 1 above, it is said that information of interest to a user can be personalized (individualized) for each user and displayed or transmitted based on the browsing history information of articles on a website. However, from the viewpoint of grasping the deep needs of each user and providing optimal information individually, it cannot be said to be sufficient, and there is room for improvement.

[0005] Therefore, the present disclosure has been made in view of the above problems, and an object thereof is to provide an information processing apparatus, an information processing method, and a program capable of accurately grasping the deep needs of each user by a new method and providing optimal information to each user.

Means for Solving the Problems

[0006] According to this disclosure, a storage unit stores product in-depth needs information that indicates the user requests that each product satisfies, associated with each of multiple products, A recommendation score calculation unit calculates a recommendation score indicating the degree of recommendation for each of the multiple products based on at least one of the user's behavioral history information and user attribute information regarding the user's actions when browsing the website. A recommendation product information determination unit that determines the ranking of products to recommend to the user on the website based on the recommendation score, An information processing device is provided, which includes a user deep needs estimation unit that estimates deep needs information indicating the desires of each user based on at least one of the recommendation score and the ranking of each product included in a predetermined number of top-ranking product groups in which the aforementioned ranking is high, and product deep needs information that is pre-associated with the product.

[0007] Furthermore, according to this disclosure, a storage process stores product in-depth needs information, which indicates the user needs that each product satisfies, associated with each of multiple products. A recommendation score calculation process that calculates a recommendation score indicating the degree of recommendation for each of the aforementioned multiple products based on at least one of the following pieces of information: behavioral history information regarding the actions of users browsing the website and user attribute information. A recommended product information determination process that determines the ranking of products to be recommended to the user on the website based on the recommendation score, An information processing method is provided in which an information processing device performs a user deep needs estimation process that estimates deep needs information indicating the desires of each user based on at least one of the recommendation score and the ranking of each product included in a predetermined number of top-ranking product groups in which the aforementioned ranking is high, and the product deep needs information that is pre-associated with the product.

[0008] Furthermore, according to this disclosure, a storage process stores product in-depth needs information, which indicates the user needs that each product satisfies, associated with each of multiple products. A recommendation score calculation process that calculates a recommendation score indicating the degree of recommendation for each of the aforementioned multiple products based on at least one of the following pieces of information: behavioral history information regarding the actions of users browsing the website and user attribute information. A recommended product information determination process that determines the ranking of products to be recommended to the user on the website based on the recommendation score, A program is provided that causes an information processing device to execute a user deep needs estimation process, which estimates deep needs information indicating the desires of each user based on at least one of the recommendation score and the ranking of each product included in a predetermined number of top-ranking products that have a high ranking, and the deep needs information of the product that is pre-associated with the product. [Effects of the Invention]

[0009] According to this disclosure, it is possible to provide an information processing device, an information processing method, and a program that can accurately grasp the deep-seated needs of each user using a new method and provide each user with the most suitable information. [Brief explanation of the drawing]

[0010] [Figure 1] This figure shows an example configuration of an information processing system according to one embodiment of the present disclosure. [Figure 2] This figure shows an example configuration of an information processing device according to the same embodiment. [Figure 3] This figure shows an example configuration of a user terminal according to the same embodiment. [Figure 4] This figure shows an example of product-specific recommendation information according to the same embodiment. [Figure 5] This figure shows an image of recommended products for the user according to the same embodiment. [Figure 6] This diagram shows the relationship between the product according to the same embodiment and the user's underlying needs. [Figure 7] This figure shows an example of output data according to the same embodiment. [Figure 8]It is a flowchart diagram related to the information processing method according to the same embodiment. [Figure 9] It is a diagram showing an example of in-depth need information for each user according to the same embodiment. [Figure 10] It is a diagram showing another example of in-depth need information for each user according to the same embodiment. [Figure 11] It is a diagram showing an example of a product in-depth need database according to the same embodiment. [Figure 12] It is a diagram showing another example of in-depth need information for each user according to the same embodiment. [Figure 13] It is a diagram showing an example of the display screen of a user terminal according to the same embodiment. [Figure 14] It is a diagram showing another example of the display screen of a user terminal according to the same embodiment.

Embodiments for Carrying out the Invention

[0011] Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In the present specification and drawings, components having substantially the same functions are denoted by the same reference numerals, and redundant descriptions are omitted.

[0012] FIG. 1 shows an example of the information processing system 100 of the present embodiment. The information processing system 100 includes an information processing apparatus 10 as a system management server and user terminals 20 used by users who are users of services using this system 100. The information processing apparatus 10 and the plurality of user terminals 20 are each connected via a network NW. The network NW is configured by the Internet, an intranet, a wireless LAN (Local Area Network), a WAN (Wide Area Network), a fixed telephone line, a mobile telephone line, etc., or a combination thereof. In this example, a plurality of users (such as site viewers, etc.) communicate with the information processing apparatus 10 via their respective user terminals 20 (20a, 20b, 20c), but it is not limited to this, and each user may use this system via the input unit and output unit of the information processing apparatus 10. That is, the present system 100 can function with only the information processing apparatus 10 without including the user terminal and the network. Alternatively, by implementing all the functions of the information processing apparatus 10 on the user terminal, it can also function independently with the user terminal.

[0013] FIG. 2 is a block diagram showing the functional configuration of the information processing apparatus 10. The information processing apparatus 10 is an apparatus used when a system administrator or the like operates and manages various services, and may be, for example, a general-purpose computer such as a workstation or a personal computer, or may be logically realized by cloud computing technology.

[0014] The information processing device 10 in this example includes a control unit 11, a storage unit 12, an input unit 13, an output unit 14, and a communication unit 15. The information processing device 10 receives various input information from the input unit 13 or from an external device via the communication unit 15. The control unit 11 then executes processing according to a program corresponding to the input information, and the processing result of the program (e.g., images or sound) is output from the output unit 14 or transmitted to an external information processing device via the communication unit 15. Note that a part of the above program may be transmitted to another information processing device (e.g., a user terminal, another server, etc.) and executed on that other information processing device. In this case, the other information processing device can be, for example, a smartphone, a mobile phone terminal, a tablet terminal, a personal computer, etc., and is connected to the information processing device 10 wirelessly or by wired connection via a network such as the Internet.

[0015] The control unit 11 handles the exchange of information between each part and controls the entire information processing device 10. This is achieved, for example, by having the CPU (Central Processing Unit), MPU (Micro Processing Unit), or GPU (Graphics Processing Unit) execute a program stored in a predetermined memory.

[0016] The storage unit 12 stores various types of information. The storage unit 12 stores programs for executing various control processes and functions within the control unit 11, input information, etc., and is composed of one or any combination thereof of RAM (Random Access Memory), ROM (Read Only Memory), flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), and other storage devices. Programs are preferably stored in a non-volatile storage medium among these storage media that constitute the storage unit 12.

[0017] The input unit 13 is for users and system administrators to input various types of information as described above, and can be implemented using, for example, a keyboard, mouse, touch panel, microphone, etc.

[0018] The output unit 14 can output information generated by the control unit 11 or information received via the communication unit. The output unit 14 may include, for example, an image output unit such as a liquid crystal display (LCD) or touch panel, and an audio output unit such as a speaker. It can output various types of information based on commands from the control unit 11. Alternatively, the output unit 14 may output (transmit) various types of information to other information processing devices (user terminals, etc.) via the communication unit 15.

[0019] The communication unit 15 is for communicating with other information processing devices and has the function of a receiving unit that receives various data and signals transmitted from other information processing devices, etc., and the function of a transmitting unit that transmits various data and signals to other information processing devices, etc., in accordance with commands from the control unit 11.

[0020] The user terminal 20 is used by users who browse websites and perform other actions using the services provided by this system 100. The user terminal 20 can be, for example, an information processing device such as a smartphone, tablet, mobile phone, or personal computer.

[0021] Figure 3 is a functional block diagram showing the user terminal 20 of Figure 1. The user terminal 20 has a control unit 21, a storage unit 22, an input unit 23, an output unit 24, and a communication unit 25. The control unit 21, like the control unit 11 described above, handles the exchange of information between each unit and controls the entire user terminal 20, and is realized, for example, by having the CPU, MPU, or GPU execute a program stored in a predetermined memory. The storage unit 22, like the storage unit 12 described above, stores programs for executing various control processes and functions within the control unit 21, input information, etc., and is composed of one or any combination thereof of RAM, ROM, flash memory, HDD, SSD, or other storage. The input unit 23, like the input unit 13 described above, is for the user to input the various information described above, and is realized, for example, by a keyboard, mouse, touch panel, microphone, etc. The output unit 24, like the output unit 14 described above, can output information generated by the control unit 21 and information received via the communication unit. For example, it may include an image output unit such as a liquid crystal display or touch panel, and an audio output unit such as a speaker. The communication unit 25, like the communication unit 15 described above, is for communicating with other information processing devices and has the function of a receiving unit that receives various data and signals transmitted from other information processing devices, and a transmitting unit that transmits various data and signals to other information processing devices in accordance with commands from the control unit 21. Users input various information via the input unit 23 of their user terminal 20 and transmit the input information to the information processing device 10, which acts as a server device, via the communication unit 25. Users also receive the processing results output from the information processing device 10 and confirm them via the output unit 24.

[0022] As shown in Figure 2, the storage unit 12 includes a product information storage unit 121, a user information storage unit 122, a recommendation information storage unit 123, and an output-related information storage unit 124. The recommendation information storage unit 123 is not a mandatory component, and the storage unit 12 may have other functional units.

[0023] The product information storage unit 121 stores various information about the product. Product information includes product identification information (product ID), product name, brand information, price information, size information, product category information, product needs information, sales status information, ranking information, review information, date and time information such as arrival date, inventory information, product image information, product description information, material information, PV count information (page views: number of times a page is viewed), etc. Product category information is information about the category (attributes) of the product, such as "shoes," "sneakers," "dresses," "sweaters," "bags," and "wallets." Sales status information is information about the sales status of the product up to the present time, including the quantity sold (number sold), total sales amount (total revenue), etc. Ranking information can be, for example, a ranking within the same brand, a ranking within the same product category, a ranking by user attribute, etc., and may be ranking information for a predetermined period (the most recent month, week, or previous day), or it may be the total for the entire period up to the present time. Ranking information can be, for example, a sales-based sales ranking, which is different from a PV count ranking. Rankings by user attribute can be by gender, age (rankings in 1-year increments), or age group (under 10, teenagers, 20s, etc., in 10-year increments, or 5-year increments, 20-year increments, etc.). Review information is information that shows the evaluation of each product. Review information may include multiple reviews. If no review is entered for a review subject, there may be no review information associated with that review subject, or only one review may be associated with it. Information entered by the user via the input unit 23 of the user terminal 20 may include text, audio, images (including still images and videos), graded evaluation values, etc. Review information entered by the user is transmitted to the information processing device 10, which acts as a server device, via the network NW and stored in the product information storage unit 121. Multiple review information may be accumulated for a single product as review information is entered by multiple users or as the same user enters review information multiple times. Date and time information may include the arrival date, as well as the expected arrival date if the product is out of stock.Furthermore, it may include information about any date or time related to the product, such as the manufacturing date. Inventory information refers to the remaining quantity (stock quantity) of the product, and product image information refers to information about images (still images or videos) related to the product.

[0024] Product-specific deep-seated needs information refers to the deep-seated needs of customers that each product can fulfill. Each product is associated with one or more pieces of deep-seated needs information. Here, a customer's "deep-seated needs" can be defined as the desires and urges hidden in the deep psychology of each customer. Users browsing websites, etc., each have their own individual deep-seated needs, and they perform actions such as selecting and deciding what to view, search for, and purchase based on these deep-seated needs. For example, if the product being viewed or purchased is "perfume," deep-seated needs might include "I want to try an expensive perfume," "I want to find a perfume that suits me," "I want to treat myself," or "I want to give it as a gift to my lover." In some cases, a user may have multiple deep-seated needs.

[0025] The user information storage unit 122 stores information about users who use this system. User information includes, for example, each user's identification information (user ID), attribute information, behavioral history information, recommendation history information, etc.

[0026] Attribute information includes, for example, each user's age, age group (including generation), gender, residential area (address), family structure, etc. Each user's attribute information may, but is not limited to, be generated based on information entered by each user. Information about each user may be automatically generated and updated based on each user's behavioral history or other information.

[0027] Behavioral history information may include online behavioral history and offline behavioral history. Online behavioral history is a history of information processing performed based on each user's input operations, and may include, for example, execution history of website browsing history, product browsing history, review viewing history, review input history, search history, purchase history (purchased products, number of purchases for each product, purchase frequency, total purchase amount over a predetermined period such as one month or one year, average purchase amount), application behavioral history, etc. Website browsing history may include, for example, site visit information for various websites visited by the user (number of site visits, site visit time, total site stay time), pages viewed by the user on each website (including articles), page viewing time (stay time) and number of views, content viewing time and number of views for content (top-ranking products, new products, campaigns, news, etc.), viewing time and number of views for each piece of content, scroll length, searched keywords, number of searches for each keyword, search frequency (number of searches per unit of time), selected categories or tags (selected category / tag information), number of times each category or tag was selected, viewed product description information, viewed review information, favorited websites, articles, and product history, selected (clicked) components within website pages (text elements including link information, image elements, etc.), etc. Product browsing history (product browsing information) may include at least one of the following: product browsing status information indicating whether or not each user viewed each product, product browsing count information indicating the number of times each product was viewed, and product browsing time information indicating the time each product was viewed. Review input history information is a record of reviews each user has entered, and includes, for example, the number of entries, frequency of entries (number of times within a specified period), date and time of entry, subject of review, review content (text, images), and a tiered evaluation parameter (numerical value). Article viewing information may include information such as the topic of the article, viewing time, number of views, and which areas were spent the most time on / clicked on. The tiered evaluation parameter is, for example, a rating on a 5-point scale, where 5 is the highest rating and 1 is the lowest, and it is a numerical value selected by the user.The memory unit 12 can store information such as the individual evaluation for each user who entered review information, and the average value for all users who entered review information for each review subject. The evaluation is not limited to a 5-point scale; it can be a 3-point scale, a 10-point scale, or any other multiple-point scale. The information processing device 10 can receive information entered by users via their user terminals, and can acquire, store, and update information regarding the history of each user's actions.

[0028] Offline activity history refers to actions taken outside of online environments, such as visiting stores or attending events. Offline activity history can be obtained, for example, by users inputting it via their user terminal, by the user terminal automatically acquiring it from usage history information of various sensors such as location sensors (GPS) or electronic money functions, by the control unit generating it based on information acquired by cameras and sensors installed in stores or event venues, or by receiving it from other external systems.

[0029] The behavioral history information preferably includes at least one of the following: search text information entered by the user, search tag information selected by the user, product viewing information related to products viewed by the user (such as product detail pages), and review viewing information related to reviews viewed by the user. The search text information entered by the user is the text information entered by the user to search for products. The search tag information selected by the user is the information of the tags (keywords) selected by the user to search for products. Information regarding whether or not a product (advertisement) suggested based on estimated deep-seated needs was selected is information regarding whether or not the suggested product was selected and the number of times it was selected.

[0030] Recommendation history information details which products (recommended items) were recommended to each user and in what order. Information about products recommended to a user is stored in association with date and time information.

[0031] The memory unit 12 stores information related to the website. This information includes text information and image information related to various pages, articles, features, and other content, as well as tags. The recommendation targets are not limited to products, but may also include services, applications, stores, websites, people, real estate, etc. Tags may include words or phrases such as spring, summer, autumn, winter, beginner, advanced, stylish, for men, for women, high price, low price, and gift, and may be associated with deep-seated product needs or deep-seated user needs.

[0032] The recommendation information storage unit 123 includes information such as conditions for determining which products to recommend. Specifically, it includes, for example, conditions for calculating the recommendation score and conditions for determining the recommendation ranking. The conditions for calculating the recommendation score can be conditions that indicate how to calculate the recommendation score based on the user's behavior history information. For example, keywords can be extracted from the articles, content, products, website titles and sentences viewed by the user, searched text, selected tags, etc., and the recommendation score can be calculated by adding recommendation scores to products related to those keywords according to their degree of relevance. The degree of relevance may be expressed by weighting the number of times each keyword has been extracted, the number of times the user has viewed the content, the viewing time, etc. Therefore, the recommendation information storage unit 123 may include condition information regarding weighting when calculating the recommendation score.

[0033] The output-related information storage unit 124 stores information related to output data. Information related to output data includes, for example, information related to output conditions and advertising information. Information related to output conditions includes, for example, condition information on what kind of screen display data to present to each user on the website based on the deep needs information of each user, and condition information on what kind of message data to output to notify or send to the user outside of the website. Information related to screen display data includes layout information on what kind of layout to arrange text and images, and condition information on what kind of text (a list of words, a sentence, selectable tags, etc.) to display and what kind of images to display based on the deep needs of the user. Information related to message data includes transmission format information such as email, push notification, or SNS distribution, and condition information on what kind of text (a list of words, a sentence, selectable tags, etc.) to use as message data and what kind of images to display. For example, when outputting text based on a user's deep-seated needs, one could apply keywords indicating those needs to a predetermined template, generate text by combining keywords based on predetermined rules (algorithms), or generate text using machine learning or generative AI with keywords as input information. The machine learning method can be unsupervised or supervised, and the method is not particularly limited as long as appropriate training data and learning algorithms are selected and executed. A trained model is obtained by machine learning a neural network using known training data and optimizing the parameters of the neural network so that desired output data is output when certain input data is input. In this example, a trained model is used to input unknown input data (user's deep-seated needs information) into the trained model to obtain predetermined output data (output text). The generation of a machine learning model may be done by training the model using training data consisting of predetermined deep-seated needs information (input data) and information on output texts corresponding to that deep-seated needs information (output data).

[0034] Advertising information stores information about advertisements as output data presented to users on a website. This information includes, for example, the products, services, applications, stores, websites, people, and real estate being advertised. Information about the advertised subject includes, for example, the name of the advertised subject, its value (price, etc.), images, review information for the advertised subject, the underlying needs associated with each advertised subject, and keywords and tags associated with each advertised subject, and is stored in association with each advertised subject. Keywords and tags include, for example, words and phrases such as spring, summer, autumn, winter, beginner, advanced, fashionable, for men, for women, high price, low price, and gift.

[0035] In this embodiment, the control unit 11 includes a recommendation score calculation unit 111, a recommended product information determination unit 112, a deep-seated user needs estimation unit 113, an output data generation unit 114, and a deep-seated product needs generation unit 115. However, the configuration of the control unit 11 is not limited to this and may include other configurations.

[0036] The recommendation score calculation unit 111 calculates a recommendation score indicating the degree to which each of several products is recommended, based on at least one of the following: behavioral history information regarding the user's actions while browsing the website and user attribute information.

[0037] The recommendation product information determination unit 112 determines the ranking of products to recommend to the user on the website based on the recommendation score. The recommendation product information determination unit 112 may decide to recommend only one product (i.e., determine only the number one product), or it may rank multiple products as recommended products. The recommendation score is a numerical value that indicates the degree to which a product is recommended to the user, and the recommendation rank is the order in which the product is recommended. Figure 4 shows an example of recommendation scores and recommendation ranks calculated for multiple products (six products in Figure 4) stored in the memory unit. In this example, the higher the recommendation score, the higher the rank, but the relationship between the recommendation score and the rank may be reversed.

[0038] Figure 5 shows an image of the recommendation ranking determined by the recommendation score calculation unit 111 and the recommendation product information determination unit 112. The recommendation score calculation unit 111 and the recommendation product information determination unit 112 determine one or more products ("recommendations for Person A") to recommend to the user (Person A) from all products stored in the memory unit in advance, and determine their rankings. Figure 5 shows rankings from 1 to 8, but rankings may be determined for all products, or rankings may be determined for only a predetermined quantity or percentage of all products.

[0039] The recommendation score calculation unit 111 may calculate recommendation scores for multiple products (all products) pre-stored in the memory unit based on behavioral history information. For example, if a user views a specific product, the system may calculate the recommendation scores of products from the same brand (brand information) as that product to be higher, resulting in products from that brand being ranked higher in recommendations. In this case, the system obtains brand information associated with the product viewed by the user from the product description, etc., and increases the recommendation scores of products from the same brand within the product group stored in the memory unit by a predetermined value (e.g., 10 points). Furthermore, within the products of that brand, the higher the popularity ranking of a product, the larger the value added to the recommendation score. Specifically, for example, 10 points may be added for 1st place, 9 points for 2nd place, 8 points for 3rd place, ... 1 point for 10th place. As a result, the recommendation scores are determined in order of popularity for products within the same brand, and the products are selected as recommended products in order of highest recommendation score.

[0040] Furthermore, the recommendation score calculation unit 111 may calculate recommendation scores for multiple products pre-stored in the memory unit based on the user's attribute information. For example, if the user is a man in his 30s, the recommendation score may be increased for products that are popular among male users in their 30s, with higher rankings receiving larger scores. The recommendation score calculation unit 111 may also use both behavioral history information and attribute information to calculate recommendation scores for multiple products pre-stored in the memory unit. In the above example, possible score calculation conditions include adding a predetermined number of scores for products of the same brand as the products viewed by the user, while adding larger scores for products that rank higher among users with the same attributes (at least one of the same gender and age group), but are not limited to these. Information referenced when calculating recommendation scores and determining recommendation rankings, as well as information regarding the conditions for score calculation and ranking determination, are pre-stored in the memory unit.

[0041] Here, the recommended products determined by the recommendation score calculation unit 111 and the recommended product information determination unit 112 may be displayed on the user terminal 20. For example, the output data generation unit 114 can display the recommended products on the user terminal 20 as recommended products in order of the highest recommendation score. More specifically, a predetermined number of products, such as the 1st to 10th ranked products, can be displayed as recommended products on the screen of the product the user is currently viewing, in order of the highest recommendation score. Information such as the recommendation score, rank, and recommended product for each product determined by the recommendation score calculation unit 111 and the recommended product information determination unit 112 may be stored in the storage unit in association with date and time information and user information. Alternatively, the calculation of the recommendation score and the determination of the recommendation rank for each product may be repeated and stored in the storage unit in association with date and time information.

[0042] The user deep needs estimation unit 113 performs a process to estimate deep needs information that indicates the desires of each user, based on at least one of the recommendation score and rank of each product included in a predetermined group of top-ranking products, and product deep needs information that is pre-associated with each product. For example, for products ranked from 1st to 10th (top-ranking product group) determined by the recommendation score, the unit refers to the memory unit and obtains all product deep needs information associated with each product. All of this information may be estimated as the user's deep needs, or only a portion of it (for example, the deep needs of a predetermined number of top products) may be estimated as the user's deep needs. Furthermore, weights may be assigned to multiple deep needs that are considered the user's deep needs. For example, a larger number may be added to the product deep needs of products with higher recommendation scores so that they rank higher as the user's deep needs. Specifically, the recommendation score may be used as is, or a value obtained by adding, multiplying, or dividing the recommendation score by a predetermined number may be added as the score for each product deep need. In addition to this, or as an alternative, higher recommendation rankings may be weighted to reflect the user's underlying needs. Specifically, the deeper the user's needs, the higher the recommendation ranking, the greater the weight given to those needs. For example, the deepest needs of the product ranked 1st might be assigned 10 points, 2nd 9 points, 3rd 8 points, ... 10th 1 point. In this way, the user's underlying needs are weighted (scored), and those with higher scores are ranked as more important. Furthermore, if some of the deepest needs associated with each product in the top product group are common (identical), weighting may be adjusted to reflect the overlap. For example, if two products, such as the 1st and 3rd ranked products, share a common deepest need, that need can be weighted twice as much to reflect the user's underlying needs more highly. The number of products designated as the top product group is predetermined and is not limited to two or more.

[0043] Figure 6 shows an example of the processing performed by the user deep needs estimation unit 113. The product ranked 1st has the deep needs "# Versatile" and "# Slimming look" pre-associated with it, the product ranked 2nd has the deep needs "# Versatile" and "# Machine washable" pre-associated with it, the product ranked 3rd has the deep need "# Versatile" pre-associated with it, the product ranked 4th has the deep needs "# Versatile" and "# Slimming look" pre-associated with it, and the product ranked 5th has the deep needs "# Slimming look" and "# Machine washable" pre-associated with it. In this case, the user deep needs estimation unit 113 determines that "# Versatile," which is the most frequent, is the deep need of the user (Person A) with the highest importance, and then determines that "# Slimming look" and "# Machine washable" are the deep needs of the next most important users. It should be noted that prioritizing these underlying user needs is not mandatory; in this case, for example, "# Versatile," "# Slimming Look," and "# Machine Washable" could all be considered the same underlying user need.

[0044] The output data generation unit 114 generates output data that includes at least one of the following, based on the deep-seated needs information of each user: screen display data to be presented to the user on the website, and message data to be notified to or sent to the user outside of the website. The output data may include, for example, advertising information for products associated with the deep-seated needs information (information about recommended products), tag information associated with the deep-seated needs information, etc. The output data may not only include the content displayed on the website, but also data output to various applications launched on the user's terminal, communication tools such as email and SNS (subject, pre-header, body, image, recommended products, etc.), or print data for objects exchanged in the real world such as advertising sheets and direct mail (body, image, recommended products, etc.). A pre-header is, for example, a short text that is not displayed in the body of an email, but is displayed only in notifications on the user's terminal, such as a smartphone or on the lock screen.

[0045] Figure 7 shows an example of output data (screen display data) that includes grouped products. In the example in Figure 7, the output data groups products according to the user's underlying needs, namely "highly versatile" and "looks neat," and displays them together on the user terminal. The output data generation unit 114 extracts all products associated with the underlying need for "highly versatile" from all products stored in the memory unit, generates a "highly versatile" group, and outputs text information of the group name (a word or sentence indicating the group name) along with information about the products in that group (various information such as images, product names, and descriptions). When generating the "highly versatile" group, the output data generation unit 114 may narrow down the selection based on predetermined conditions. For example, it may exclude products that are not suitable for the user's attributes from among the products associated with the underlying need for "highly versatile." More specifically, if the user is male, it may limit the selection to men's products, or if the user's size is "L," it may limit the selection to L-size products.

[0046] In the example in Figure 7, the output data generation unit 114 can generate a group called "Clean Look" and present the products in that group together. The products displayed as output data may or may not include the original top-recommended product group. The output data may also consist of 10 products that make up the top-recommended product group, sorted in order of importance to the user's deep-seated needs. The order in which the groups are displayed as output data may be in descending order of the deep-seated needs score for each user, or it may not be. In this example, the "Highly Versatile" group, which has high importance, is displayed in a higher position (at the top in this example, but not limited to this). Furthermore, the order in which products are displayed within each group may be such that those with higher importance to the deep-seated needs for each user are placed higher, or they may be in other orders such as ranking order, order of arrival date (closest or farther from the current date), order of identification number, or alphabetical order.

[0047] Figure 8 is a flowchart showing an example of an information processing method performed by this system. The information processing method includes deep product needs storage processing (S1), recommendation score calculation processing (S2), recommended product information determination processing (S3), user deep needs estimation processing (S4), and output data generation processing (S5), all of which are performed by the control unit 11.

[0048] In S1, the control unit 11 stores, for example, input information from the system administrator, all products stored in the memory unit in association with product deep needs information that indicates the user's requirements that each product fulfills. The product deep needs information may be information entered by the system administrator, or it may be generated by the product deep needs generation unit 115, or modified (updated) by the product deep needs generation unit 115.

[0049] In S2, the recommendation score calculation unit 111 calculates a recommendation score indicating the degree of recommendation for each of several products based on at least one of the following: behavioral history information regarding the user's actions while browsing the website and user attribute information.

[0050] In S3, the recommended product information determination unit 112 determines the ranking of products to recommend to users on the website based on the recommendation score.

[0051] In S4, the user deep needs estimation unit 113 estimates deep needs information that indicates the desires of each user, based on at least one of the recommendation score and rank of each product included in a predetermined number of top-ranking product groups, and product deep needs information that is pre-associated with the product.

[0052] In S5, the output data generation unit 114 generates output data that includes at least one of the following: screen display data to be presented to the user on the website, and message data to be notified to or sent to the user outside of the website, based on the deep-seated needs information of each user.

[0053] The processes from S1 to S5 may be executed repeatedly, either in whole or in part. In this case, the estimation of deep-seated user needs and the generation of output data can be repeated at predetermined intervals, such as 5 seconds, 10 seconds, 30 seconds, 1 minute, or 5 minutes. When the process is repeated at short intervals, such as 5 seconds or 10 seconds, personalized output data can be presented to the user in a near real-time state, depending on the user's behavioral history information. Alternatively, the calculation of recommendation scores and the ranking of recommended products may be performed every 12 hours, 24 hours (1 day), or 1 week. Furthermore, so-called ad-hoc execution is also permitted. For example, by performing at least one of the calculation of recommendation scores and the ranking of recommended products each time behavioral data is acquired, and updating the information, it is possible to ensure real-time responsiveness to user behavior. Specifically, for example, if a user views product A, there is a high probability that they are viewing products of the same brand and category as product A, so the recommendation score for products related to (similar to) product A can be reflected more highly.

[0054] As described above, in this embodiment, based on the product recommendation results (recommendation score, recommendation ranking) for the user, we estimate the deep-seated needs information that the user is likely to be concerned with. In conventional recommendations, it was possible to determine and present recommended "products" from the user's information, but it was not possible to appeal to the user deeply through linguistic information. Specifically, it was limited to abstract phrases such as "Recommended for you" or "Also recommended." Furthermore, methods that simply group and display the features of multiple products to be promoted (for example, assigning a cardigan to the slogan "A cardigan is a must-have for autumn with drastic temperature changes") allow for the creation of deeper appeals, but are insufficient in terms of being suitable for individual users. In this embodiment, first, recommended products are extracted for each user (by a predetermined recommendation algorithm, etc.), and common features are extracted from that group of recommended products as "deep-seated needs," so that these "deep-seated needs" have a higher probability of appealing to the deep psychology of individual users. Furthermore, by also estimating products linked to those underlying needs, it becomes possible to recommend products appropriately with a deeper appeal. As described above, this new method makes it possible to accurately grasp each user's underlying needs and provide personalized and optimal information to each user.

[0055] Furthermore, in this embodiment, by combining recommended product information with deep-seated product needs information (database), it is possible to estimate and articulate the customer's deep-seated needs, thereby achieving personalization. While there have been conventional services that simply personalize based on behavioral data, they cannot explain "why that personalization was performed? Why is that product being recommended?" In this embodiment, by combining behavioral history information with deep-seated product needs information, the underlying intention can be interpreted and articulated in the form of deep-seated needs. As a result, personalization actions beyond simply recommending products can be performed (for example, searching by needs tags, creating captions (headlines, titles, descriptions) for recommendations using deep-seated needs, inserting phrases that resonate with deep-seated needs into appealing messages such as emails), thereby increasing the breadth and depth (accuracy) of personalization.

[0056] The user deep needs estimation unit 113 may also estimate the category and importance of each user's requests. For example, in the top group of products, category information may be extracted from the deep needs information of the products associated with the products, in order of highest recommendation score (or in order of highest recommendation ranking), and the user's deep needs may be estimated such that the higher the recommendation score, the higher the importance of that category. In this case, in determining importance, a value based on the recommendation score may be added (multiplied, subtracted, or divided), or a predetermined value based on the recommendation ranking may be added (multiplied, subtracted, or divided). Figures 9 and 10 show examples of categories and importance as deep needs information for each user, respectively. Thus, the method of representing user deep needs information is not limited to one form.

[0057] In this embodiment, the output data generation unit 114 may generate product groups consisting of multiple products that satisfy common needs, based on the deep-seated needs information of the user estimated by the deep-seated needs estimation unit 113, and output screen display data or message data related to the product groups. With such a configuration, it is possible to present the user with product groups that are even more accurate in meeting the user's needs.

[0058] In this embodiment, the output data generation unit may output screen display data or message data in which multiple products are arranged in an order different from the recommendation ranking determined by the recommendation product information determination unit, based on the deep-seated needs information of the user estimated by the deep-seated needs estimation unit. With such a configuration, it is possible to present products to the user in an order that is more accurate in line with the user's requests, rather than simply arranging them in order of recommendation score.

[0059] In this embodiment, the behavioral history information includes product viewing information, which includes at least one of the following: product viewing status information (whether or not each user viewed each product), product viewing count information (how many times each product was viewed), and product viewing time information (how long each product was viewed). The recommendation product information determination unit may add a recommendation score for products related to the products viewed by the user based on the product viewing information. For example, depending on whether or not the user viewed each product, a predetermined value may be added if the user viewed it, and not added if the user did not view it. Alternatively, a predetermined value may be added if the product viewing count information is equal to or greater than a predetermined number, and not added if it is less than a predetermined number. Alternatively, a predetermined value may be added if the product viewing time is equal to or greater than a predetermined time, and not added if it is less than a predetermined time. Or, a combination of these conditions may be used to calculate the recommendation score.

[0060] In this embodiment, the recommended product information determination unit 112 may calculate a product recommendation score based on the user's attribute information and at least one of the sales status information, evaluation information, and ranking information of the product among a group of users with the same attributes as the user. For example, if the number of sales as sales status information for the product exceeds a predetermined value, a predetermined value may be added. If the average value of reviews as evaluation information (e.g., 3.5 on a 5-point scale) is above a predetermined value, a predetermined value may be added. Also, if the ranking is above a predetermined rank, a predetermined value may be added. Alternatively, a combination of several of these conditions may be used to calculate the recommendation score.

[0061] The User Deep Needs Estimation Unit 113 may further estimate the deep needs of each user based on their behavioral history information. For example, it may estimate the deep needs of users who have viewed articles such as "Special Feature on Sale Items" and whose viewing information meets predetermined conditions (such as users whose viewing count, viewing time, and viewing frequency are above predetermined values) as "highly sensitive to sales," or it may estimate the deep needs of users whose viewing information for articles such as "Special Feature on Stylish Functional Outerwear for Winter" meets predetermined conditions as "wanting outerwear" and "wanting both functionality and style." In this case, the deep needs may be estimated by extracting keywords (such as "sale," "winter," "functionality," and "outerwear") from text information associated with viewed articles, viewed pages, and viewed products as behavioral history information, and generating sentences based on predetermined conditions by combining one or more keywords. When estimating deep needs, the User Deep Needs Estimation Unit 113 may use a generative AI such as GPT (Generative Pretrained Transformer) to generate sentences representing deep needs.

[0062] Furthermore, the information processing device of this embodiment may include a deep-root product needs generation unit that generates deep-root product needs information for each of the products based on input information, including review information entered by users for products listed on the website. In this case, for example, by generating deep-root product needs based on the opinions (review information) of customers who evaluate the products, rather than the opinions of the companies that sell the products, it becomes possible to grasp the needs that the products satisfy from a customer's perspective, which is one level deeper than the existing understanding of products by the companies.

[0063] Figures 9 and 10 show examples of deep-seated user needs estimated by the deep-seated user needs estimation unit 113. Each user of the website is assigned identification information (such as a personal ID) by the control unit 10. Such identification information is unique to each user and is expressed as a combination of numbers, letters, and other characters. Based on each user's recommended product information and product deep-seated needs information, the deep-seated needs of one or more users and the importance of each deep-seated need can be determined and stored in the memory unit.

[0064] In the example in Figure 9, each user's database stores user identification information, attribute information such as "gender" and "age," and data on deep-seated needs, including items such as "emphasis on functionality," "for going out," "categories of interest," and "interest in specific elements." The "emphasis on functionality" item indicates the degree to which the user values ​​the functionality of the product, and is classified into three levels: "high," "medium," and "low," from highest to lowest. The "for going out" item indicates the degree to which the user wants to use the product when going out, and is classified into four levels: "high," "medium," and "low," plus "-" (hyphen) for unknown. Note that the levels for all items can be changed as appropriate; for example, there could be two levels such as "yes" and "no" or "high" and "low," or there could be four or more levels with more detailed classifications. For example, the degree of desire for deep-seated needs could be expressed using a numerical value between 0 and 100. The "categories of interest" item contains information on specific product categories (classifications) that each user is particularly interested in (desiring), including information such as "pumps" and "yoga mats." The "Interest in Specific Elements" item is information that classifies the degree of interest in a particular brand, category, product, or store on a three-point scale, for example. Furthermore, if there are multiple deep-seated needs items such as "Functionality-Focused" or "For Outings," the user deep-seated needs estimation unit 113 may estimate and store their priority order. For example, the priority order of four items—"Functionality-Focused," "For Outings," "Featured Categories," and "Interest in Specific Elements"—may be estimated from the user's behavioral history information. In this case, for example, the priority order might be "Functionality-Focused," "For Outings," "Featured Categories," and "Interest in Specific Elements." This allows, for example, the output data generation unit 114 to prioritize selecting advertisements related to the highest-priority item as output data.

[0065] As illustrated in Figure 10, the memory unit may store user identification information (personal ID) and information constituting deep-seated needs, such as "category," "needs (detailed requests)," and "importance," in association with each user. Alternatively, the information constituting deep-seated needs, such as "category," "needs," and "importance," may be stored in the memory unit as pre-selected options, and the user deep-seated needs estimation unit 113 may select the appropriate deep-seated needs for each user from these options based on each user's behavioral history information. Alternatively, the user deep-seated needs estimation unit 113 may generate individual information such as "category," "needs," and "importance" that constitute deep-seated needs based on each user's behavioral history information and store it in the memory unit. Furthermore, output data such as advertising information may be pre-associated and stored with each deep-seated need. The category is the field to which the requested object belongs, and the needs are the specific requests. Importance is indicated in multiple stages, such as three levels: "high," "medium," and "low," to show the priority of the deep-seated needs.

[0066] Furthermore, the user deep needs estimation unit 113 may estimate each user's deep needs not only based on the user's own information but also on information from others. For example, it can estimate each user's deep needs by referring to the recommended product information, behavioral history information, deep needs information, etc., of others that have a high similarity to the user's behavioral history information. The user deep needs estimation unit 113 may estimate the deep needs of the other user with the highest similarity as the user's own deep needs, or it may estimate the user's own deep needs based on the deep needs of multiple other users with a high similarity. The user deep needs estimation unit 113 can calculate the similarity of site browsing history information between one user and another user.

[0067] The user deep needs estimation unit 113 can compare one user's site browsing history information with another user's site browsing history information to determine the similarity and store it in the memory unit. The similarity is expressed as a numerical value (%) from 0 to 100, for example, with a higher number indicating a higher degree of similarity. Similarity is determined by calculating a similarity score by comparing one or more combinations of any of the components included in the site browsing history information, such as the object of browsing (articles, products, etc.), display time (time for one time, total of multiple times, average, etc.), display time, number of times displayed, tags, keywords, etc., and the number of times they are selected. The similarity score can be calculated by adding a predetermined value when there is one instance of commonality, such as browsing the same website, product, article, etc., matching selected tags, or matching search keywords. The method for determining similarity is not particularly limited, but for example, the number of times the same tag was selected, the number of times the same product was viewed, or the number of times the same page on a website was viewed may be compared, and the user with the smallest difference may be determined. The parameters to be compared are not limited to the number of views mentioned above; they may also include the length of time spent viewing the same product, page, etc. Viewing means displaying the item on the screen. The determination or calculation of similarity may be based on machine learning using a learning model or other analytical methods.

[0068] The control unit 11 in this example has a deep-rooted product needs generation unit 115. The deep-rooted product needs generation unit 115 generates deep-rooted product needs information associated with a product based on input information related to the product and stores it in a storage unit (product information storage unit 121). The deep-rooted product needs generation unit 115 may generate a database of deep-rooted needs for products, etc., based on machine learning using a learning model or other analytical methods. In that case, a trained model may be generated by first training a combination of information including text and images (input information) such as product descriptions, product names, brand names, titles of articles in which the product is featured, article text, website titles, content titles, etc., and information on the deep-rooted needs that each product satisfies (output information) as training data. The deep-seated product needs generation unit 115 takes data related to the product, such as metadata, product descriptions, product review information, and product images, as input information, and analyzes it using a predetermined algorithm or machine learning to extract (estimate) the deep-seated needs that the product fulfills (for example, "pumps A," "comfortable to wear," "perfect for a date"). Specifically, for example, in the case of a product featured in an article titled "Featured Highly Functional and Stylish Sneakers," the deep-seated needs such as "functionality" and "stylish" can be estimated. As input information, the unit may also acquire information about the product (metadata, product descriptions, product review information, product images) from external websites or external systems such as social media, including review information entered by users for products posted on the website, and generate deep-seated product needs information based on these different websites.

[0069] Figure 11 shows an example of a product deep-seated needs database. Each product is associated with deep-seated needs information that it fulfills ("cleanliness," "elegance," "sophistication," "gentleness," "floral"). As shown in Figure 11, each product is not limited to fulfilling just one deep-seated need; it may be multiple. In the example in Figure 11, each product's identification information is associated with one or more deep-seated needs. Such a database may be pre-configured by the service provider (system administrator). When the product deep-seated needs generation unit 115 generates the database, for example, it may estimate the deep-seated needs pre-associated with users who have purchased each product as the deep-seated needs corresponding to that product, or it may estimate the deep-seated needs corresponding to a product based on deep-seated needs estimated based on at least one of the user's site browsing history and review viewing history, and data such as the user's viewing time, number of searches, and number of tag selections, and then generate the database.

[0070] Figure 12 shows an example of a database associated with deep-seated needs for each user (customer). This database can be generated by associating deep-seated needs estimated based on each user's website browsing history, recommended product information, etc., with each user. Furthermore, the user's deep-seated needs database can also be generated based on a product-related deep-seated needs database, as shown in Figure 11, and information about each user (attribute information, purchase information, behavioral history information, or a combination thereof). The control unit can output the generated database as output data, which can then be displayed, for example, on a user terminal's touch panel, monitor, or other display device. Such a database can be provided, for example, to product vendors, enabling them to offer product suggestions tailored to the user's deep-seated needs.

[0071] The output data generation unit 114 generates output information (output data) to be provided to the user based on various input information. For example, the output data generation unit 114 generates output data to be output to each site visitor based on the deep needs information of each site visitor estimated by the deep needs estimation unit 113. The output data may include, for example, advertising information for products associated with the deep needs information (information about recommended products), tag information associated with the deep needs information, etc. The output data may not only be the content displayed on the website, but also data output to various applications launched on the user's terminal, communication tools such as email and SNS (subject, preheader, body, image, recommended products, etc.), or print data for objects exchanged in the real world such as advertising sheets and direct mail (body, image, recommended products, etc.). A preheader is, for example, a short text that is not displayed in the body of an email, but is displayed only in notifications on the user's terminal such as a smartphone or on the lock screen.

[0072] The output data generation unit 114 generates output data such as advertisements based, for example, on one or more deep-seated needs items (categories or needs) for each user and their degree (importance). For example, for a user with ID number "0001", it can generate one or more advertisements as output data related to pumps of a specific brand that are highly functional, stored as suitable for going out, and have specific characteristics.

[0073] Furthermore, the output data generation unit 114 can determine which advertising information to prioritize based on the estimated deep-seated user needs and, if necessary, the importance of each item. In other words, the output data generation unit 114 can set priorities by displaying advertisements that address high-priority deep-seated needs earlier in time than advertisements that address low-priority deep-seated needs, or by placing them in a position where they are displayed first (e.g., at the top when scrolling from top to bottom, or on the left when scrolling from left to right).

[0074] Figures 13 and 14 show examples of user-specific output data displayed on a website. The screen displaying the output data is not limited to a shop screen where products are listed, but can be any screen. For example, it could be a screen that displays one or more items in the cart that are temporarily stored as items under consideration for purchase (a screen to confirm items added to the cart). Specifically, by selecting an icon such as "Add to Cart" associated with any item under consideration for purchase, it is stored as an item under consideration for purchase. Then, in response to input such as selecting an icon that instructs the user to display the items in the cart, a screen to confirm the items added to the cart is displayed, and advertisements for related products can be displayed on that screen as output data.

[0075] In the example in Figure 13, information about multiple specifically grouped products (pumps) (images, prices, product names, brands, etc.) is displayed from left to right in order of priority, along with text (composite information) containing category information for the deep-seated need, "Recommended pumps that are comfortable and suitable for summer dates!", which corresponds to the most important estimated deep-seated user need information. Note that the text (caption, heading, title, description) as composite information for deep-seated needs may be generated by combining individual elements of the deep-seated need ("summer," "date," "functionality," "pumps," etc.), or it may be generated from only one individual element of the deep-seated need (for example, the text "For those who love limited editions!" may be generated from only the deep-seated need "love of limited editions"). If there are multiple suggested products, the priority of the suggested products (advertisements) as output data is determined in the order in which they satisfy the most individual elements of the deep-seated need. Furthermore, in the example in Figure 13, multiple specific products (skirts) are displayed from left to right in order of priority, along with category information for the next most important deep-seated user need information, "Stylish skirts recommended for summer dates." In this case, the individual underlying needs of the user are "summer," "date," "elegant," and "skirt."

[0076] In the example shown in Figure 14, multiple tag information ("#Doesn't hurt your feet", "#Perfect for summer dates", "#Easy to match", "#Breathable") associated with the estimated deep-seated user needs are displayed as selectable options. By selecting any of the tags, the user can view products associated with that tag. The output data generation unit 114 can generate one or more such tag information as output data.

[0077] The output data generation unit 114 may further generate output data based on user attribute information, such as that of site visitors. Attribute information may include, for example, age, gender, residential area (address), and family structure. This allows for the provision of more accurate output data tailored to each user's attribute information.

[0078] The output data generation unit 114 may generate output data to highlight the advertisement text on individual product pages. Highlighting can, but is not limited to, making the font size larger than other text, changing the color, making the text outline thicker, or underlining the text. The output data generation unit 114 may also automatically display recommended fields, review subjects (products), review information, etc., in a pop-up window.

[0079] Furthermore, in the information processing device of this embodiment, the user deep needs estimation unit may also estimate the deep needs information of each user based on the attribute information of each user. This makes it possible to estimate deep needs information based on attribute information such as the user's age and place of residence. As a result, it becomes possible to more accurately grasp the deep needs of users and provide appropriate information.

[0080] Furthermore, in the information processing device of this embodiment, the output data generation unit may also generate output data to be output to each user based on the attribute information of each user. This allows output data to be generated based on attribute information such as the user's age and place of residence. As a result, appropriate information according to the user's attributes can be provided.

[0081] In this embodiment, the output data may include the email title (subject) and body text. That is, the output data generation unit 114 can determine the title and body text of the email to be sent to each user based on the user's deep-seated needs. This makes it possible to automatically create and send emails to email addresses stored in the memory unit in advance through user registration, etc. As a result, it is possible to send emails with content tailored to the deep-seated needs of each individual user, making it easier to capture the user's attention compared to emails with uniform text. Furthermore, the title and body text of the email may include keywords used as tag information in the previous embodiment, or text tailored to the deep-seated needs may be generated. For the generation of specific text, machine learning techniques (including generative AI) using a learning model may be used.

[0082] While preferred embodiments of the present disclosure have been described in detail above with reference to the attached drawings, the technical scope of the present disclosure is not limited to such examples. It is clear to any person with ordinary skill in the art of the present disclosure that various modifications or alterations may be conceived within the scope of the technical idea set forth in the claims, and these will naturally also fall within the technical scope of the present disclosure.

[0083] The apparatus described herein may be implemented as a single device, or it may be implemented as a group of devices (e.g., cloud servers) that are partially or entirely connected by a network. For example, the control unit 11 and the storage unit 12 of the information processing apparatus 10 may be implemented as different servers connected to each other by a network.

[0084] The series of processes performed by the apparatus described herein may be implemented using software, hardware, or a combination of software and hardware. Computer programs for implementing each function of the information processing apparatus 10 according to this embodiment can be created and implemented on a PC or the like. Furthermore, a computer-readable recording medium containing such a computer program can also be provided. Examples of recording media include magnetic disks, optical disks, magneto-optical disks, and flash memory. Alternatively, the computer program may be distributed without using a recording medium, for example, via a network.

[0085] Furthermore, the processes described using flowcharts in this specification do not necessarily have to be executed in the order shown. Some processing steps may be executed in parallel. Additional processing steps may be adopted, and some processing steps may be omitted.

[0086] Furthermore, the effects described herein are merely descriptive or illustrative and not limiting. In other words, the technology relating to this disclosure may produce other effects that will be apparent to those skilled in the art from the description herein, in addition to or in lieu of the effects described herein.

[0087] Furthermore, the following configurations also fall within the technical scope of this disclosure. (Item 1) A memory unit that stores deep-seated product needs information, which indicates the user requests that each product fulfills, associated with each of multiple products. A recommendation score calculation unit calculates a recommendation score indicating the degree of recommendation for each of the multiple products based on at least one of the user's behavioral history information and user attribute information regarding the user's actions when browsing the website. A recommendation product information determination unit that determines the ranking of products to recommend to the user on the website based on the recommendation score, An information processing device comprising: a user deep needs estimation unit that estimates deep needs information indicating the desires of each user based on at least one of the recommendation score and the ranking of each product included in a predetermined number of top-ranking product groups in which the aforementioned ranking is high, and product deep needs information pre-associated with the product. (Item 2) The user deep needs estimation unit is an information processing device as described in item 1, which estimates the category and importance of each user's request. (Item 3) The information processing apparatus according to item 1, further comprising an output data generation unit that generates output data including at least one of screen display data to be presented to the user on the website and message data to be notified to or sent to the user outside of the website, based on the deep-seated needs information of each user. (Item 4) The output data generation unit generates a product group consisting of multiple products that satisfy the common requirements, based on the deep-seated user needs information estimated by the deep-seated user needs estimation unit. The information processing device described in item 2, which outputs the screen display data or message data relating to the aforementioned product group. (Item 5) The information processing device according to item 2, wherein the output data generation unit outputs screen display data or message data in which multiple products are arranged in an order different from the recommendation ranking by the recommendation product information determination unit, based on the deep needs information of the user estimated by the deep needs estimation unit of the user. (Item 6) The aforementioned behavioral history information includes product viewing information, which includes at least one of the following: product viewing status information indicating whether or not each user viewed each product; product viewing count information indicating the number of times each product was viewed; and product viewing time information indicating the time each product was viewed. The information processing device described in item 1, wherein the recommended product information determination unit adds recommendation scores for products related to the products viewed by the user, based on the product viewing information. (Item 7) The information processing device described in item 1, wherein the recommended product information determination unit calculates a product recommendation score based on the user's attribute information and at least one of the sales status information, evaluation information, and ranking information of products in a group of users with the same attributes as the user. (Item 8) A memory processing method that stores deep-seated product needs information, which indicates the user needs that each product fulfills, associated with each of multiple products. A recommendation score calculation process that calculates a recommendation score indicating the degree of recommendation for each of the aforementioned multiple products based on at least one of the following pieces of information: behavioral history information regarding the actions of users browsing the website and user attribute information. A recommended product information determination process that determines the ranking of products to be recommended to the user on the website based on the recommendation score, An information processing method comprising: an information processing device performing a user deep needs estimation process that estimates deep needs information indicating the desires of each user based on at least one of the recommendation score and the ranking of each product included in a predetermined number of top-ranking product groups in which the aforementioned ranking is high, and the product deep needs information that is pre-associated with the product. (Item 9) A memory processing method that stores deep-seated product needs information, which indicates the user needs that each product fulfills, associated with each of multiple products. A recommendation score calculation process that calculates a recommendation score indicating the degree of recommendation for each of the aforementioned multiple products based on at least one of the following pieces of information: behavioral history information regarding the actions of users browsing the website and user attribute information. A recommended product information determination process that determines the ranking of products to be recommended to the user on the website based on the recommendation score, A program that causes an information processing device to execute a user deep needs estimation process that estimates deep needs information indicating the desires of each user, based on at least one of the recommendation score and the ranking of each product included in a predetermined number of top-ranking product groups that have a high ranking, and the deep needs information of the product that is pre-associated with the product. [Explanation of Symbols]

[0088] 100 Information Processing Systems 10 Information Processing Devices 20 User Terminals

Claims

1. A memory unit that stores deep-seated product needs information, which indicates the user requests that each product fulfills, associated with each of multiple products. A recommendation score calculation unit calculates a recommendation score indicating the degree of recommendation for each of the multiple products based on at least one of the user's behavioral history information and user attribute information regarding the user's actions when browsing the website. A recommendation product information determination unit that determines the ranking of products to recommend to the user on the website based on the recommendation score, An information processing device comprising: a user deep needs estimation unit that estimates deep needs information indicating the desires of each user based on at least one of the recommendation score and the ranking of each product included in a predetermined number of top-ranking product groups in which the aforementioned ranking is high, and product deep needs information pre-associated with the product.

2. The information processing apparatus according to claim 1, wherein the user deep needs estimation unit estimates the category and importance of each user's request.

3. The information processing apparatus according to claim 1, further comprising an output data generation unit that generates output data including at least one of screen display data to be presented to the user on the website and message data to be notified to or sent to the user outside of the website, based on the deep-seated needs information of each user.

4. The output data generation unit generates a product group consisting of multiple products that satisfy the common requirements, based on the deep-seated user needs information estimated by the deep-seated user needs estimation unit. The information processing apparatus according to claim 2, which outputs the screen display data or message data relating to the product group.

5. The information processing apparatus according to claim 2, wherein the output data generation unit outputs screen display data or message data in which multiple products are arranged in an order different from the recommendation ranking by the recommendation product information determination unit, based on the deep needs information of the user estimated by the deep needs estimation unit.

6. The aforementioned behavioral history information includes product viewing information, which includes at least one of the following: product viewing status information indicating whether or not each user viewed each product; product viewing count information indicating the number of times each product was viewed; and product viewing time information indicating the time each product was viewed. The information processing device according to claim 1, wherein the recommendation score calculation unit adds recommendation scores for products related to the products viewed by the user based on the product viewing information.

7. The information processing apparatus according to claim 1, wherein the recommendation score calculation unit calculates a recommendation score for a product based on the user's attribute information and at least one of the sales status information, evaluation information, and ranking information of the product among a group of users with the same attributes as the user.

8. A memory processing method that stores deep-seated product needs information, which indicates the user needs that each product fulfills, associated with each of multiple products. A recommendation score calculation process that calculates a recommendation score indicating the degree of recommendation for each of the aforementioned multiple products based on at least one of the following pieces of information: behavioral history information regarding the actions of users browsing the website and user attribute information. A recommended product information determination process that determines the ranking of products to be recommended to the user on the website based on the recommendation score, An information processing method comprising: an information processing device performing a user deep needs estimation process that estimates deep needs information indicating the desires of each user based on at least one of the recommendation score and the ranking of each product included in a predetermined number of top-ranking product groups in which the aforementioned ranking is high, and the product deep needs information that is pre-associated with the product.

9. A memory processing method that stores deep-seated product needs information, which indicates the user needs that each product fulfills, associated with each of multiple products. A recommendation score calculation process that calculates a recommendation score indicating the degree of recommendation for each of the aforementioned multiple products based on at least one of the following pieces of information: behavioral history information regarding the actions of users browsing the website and user attribute information. A recommended product information determination process that determines the ranking of products to be recommended to the user on the website based on the recommendation score, A program that causes an information processing device to execute a user deep needs estimation process that estimates deep needs information indicating the desires of each user, based on at least one of the recommendation score and the ranking of each product included in a predetermined number of top-ranking product groups that have a high ranking, and the deep needs information of the product that is pre-associated with the product.

Images