JP6698575B2 - Recommendation system and recommendation method - Google Patents

Description

  The present invention relates to a recommendation system and a recommendation method.

  Currently, the development of a system using a technology called AI (Artificial Intelligence) is being actively conducted. As a system using AI, for example, there is a system that makes a recommendation (recommendation) according to a user's hobbies or preferences (see, for example, Patent Document 1), a system that performs image recognition on an input image and video, and the like. Are known.

  Further, at present, there is a technology capable of performing more advanced information processing by exchanging information between all things (objects) through a communication network such as the Internet, as called IoT (Internet of Things). Proposed.

JP 2009-252177

  As in the technique described in Patent Document 1, in the conventional recommendation system, one system is closed to perform the recommendation. For example, in the case of a system that recommends according to the user's hobbies and preferences, the system recommends (recommends) according to the user's hobbies and preferences from the information (for example, sales products) managed in the system. I was selecting products.

  However, given that many systems using AI technology are currently provided, the systems using these AI technologies work together to enable more appropriate recommendations while being personalized to the user. It is considered to be.

  Therefore, an object of the present invention is to provide a technique capable of performing personalized recommendations by linking a plurality of systems using the AI technique.

  A recommendation system according to an aspect of the present invention is a recommendation system including a first device used by an unspecified user and a second device used by a specified user, wherein the first device is a first device. A first recommendation control unit configured to perform recommendation based on first attribute information managed by the device, and notify the second device of the recommendation information including the content of the recommendation, and the second device notifies from the first device. A second recommendation control unit configured to notify the first device of response information based on the recommended information and the second attribute information related to the specific user, and the first recommendation control unit includes the recommendation information based on the response information. Notifies the second device of other recommendation information including the content of a different recommendation.

A recommendation method according to an aspect of the present invention is a recommendation method executed by a recommendation system including a first device used by an unspecified user and a second device used by a specific user, the first method comprising: The device makes a recommendation based on the first attribute information managed by the first device, and notifies the second device of the recommendation information including the content of the recommendation, and the second device is notified by the first device. Notifying the first device of response information based on the recommendation information and the second attribute information regarding the specific user,
The first device notifies the second device of other recommendation information including the content of the recommendation different from the recommendation information based on the response information.

  According to the present invention, it is possible to provide a technique capable of performing personalized recommendations by linking a plurality of systems using the AI technique.

It is a figure showing an example of a system configuration of a recommendation system concerning this embodiment. It is a figure which shows an example of the functional block structure of the user terminal which concerns on this embodiment. It is a figure which shows an example of the functional block structure of the IoT apparatus which concerns on this embodiment. It is a sequence diagram which shows an example of the operation|movement outline|summary which the recommendation system which concerns on this embodiment performs. It is a figure for explaining the example 1 of operation. It is a figure for explaining the example 2 of operation.

  A preferred embodiment of the present invention will be described with reference to the accompanying drawings. In addition, in each of the drawings, components denoted by the same reference numerals have the same or similar configurations.

<System configuration>
FIG. 1 is a diagram showing an example of the system configuration of the recommendation system according to the present embodiment. The recommendation system according to the present embodiment includes an IoT device 10 (first device) and a user terminal 20 (second device). The IoT device 10 and the user terminal 20 are connected via a communication network N so that they can communicate with each other. Although one IoT device 10 and one user terminal 20 are shown in FIG. 1, two or more IoT devices 10 and user terminals 20 may be included in each.

  For the communication network N, for example, a mobile radio communication technique called 3G, LTE (Long Term Evolution) or 5G may be used, or a radio communication technique such as a wireless LAN or Bluetooth (registered trademark) may be used. Good. Alternatively, wireless and wired communication may be combined.

  The IoT device 10 is assumed to be a terminal used by an unspecified user, and has a function of making various recommendations based on attribute information managed by the IoT device 10. An example of such an IoT device 10 is, for example, an IoT device 10 installed in a store and recommending recommended products to customers who visit the store. Such attribute information managed by the IoT device 10 is, for example, information on each product sold at the store where the IoT device 10 is installed (information such as product name and price, attribute of product (color and type). Etc.), information on top-selling products, etc.).

  In the present embodiment, the IoT device 10 includes any device having a communicable function, but may be, for example, a device having a smart window capable of displaying arbitrary information. The smart window includes, for example, a transmissive display, a display in which a window or glass is combined with a film and a projector, an ultra high definition (eg UHD (Ultra High Definition)) display, a non-transmissive display, or the like. Good. A transmissive display is not limited to one that directly transmits light, for example, a camera installed on the back side of the display, and it is a display that looks like it is transmitting by displaying the image taken by the camera on the display. It may be. Such smart windows are, for example, train windows, aircraft windows, ship windows, car windows, bus windows, aquarium glasses, shop windows, showcases, museum glasses, museum glasses, private house windows. The IoT device 10 according to the present embodiment may be configured by being incorporated in all windows such as windows, windows of observation facilities, and windows of buildings, and glass.

  The user terminal 20 is a terminal used by the user and has a function of making various recommendations (recommendations) for the user based on attribute information about the user. The user terminal 20 is a portable terminal that can be carried by the user, such as a smartphone, a tablet terminal, or a notebook PC (Personal Computer). Since the IoT device 10 is a terminal used by an unspecified user, the user terminal 20 can be defined as a terminal used by a specific user for the purpose of comparison with this.

  The attribute information regarding the user is information regarding the attribute of the user, and includes, for example, the basic attribute of the user, the attribute indicating the behavior of the user, the attribute indicating the family relationship of the user, the attribute indicating the friendship relationship of the user, and the preference of the user. And attributes relating to the user's living body. The basic attributes of the user are, for example, the user's name, sex, age, blood type, occupation, company of work, country of origin, language used, time zone of country of origin, and the like. The attribute indicating the user's action is, for example, a movement log (history of location information), an action schedule (when, where, what was done/what to do, what is being done now), purchase history ( Where and what was bought), website access log, website search log, etc. The attribute indicating the user's family relationship is, for example, information about the family (names and ages of parents and children) and the like. The attribute indicating the friendship relationship of the user is, for example, information on the friend composition, information indicating the business friendship relationship, or the like. The attributes indicating the user's preference include, for example, preference regarding food (favorite food, food that is poor), preference regarding color, preference regarding clothing (type of favorite coat, jeans rather than skirt, etc.), hobby, etc. is there. The attributes relating to the user's living body are, for example, the posture of the user, average calorie consumption, activity amount, sleep condition, heart rate, blood pressure, pulse rate, and the like.

  In the present embodiment, the IoT device 10 and the user terminal 20 cooperate with each other to make it possible to perform more appropriate recommendation according to the user's TPO (Time Position Occasion) and the like while being personalized by the user.

  In addition, TPO is a term meaning the proper use according to time, place, and case. Examples of TPO when the user is moving include sightseeing, work, honeymoon, family vacation, traveling alone, homecoming, shopping. It's a transfer, a commute, a drive, a legal affair, a premarital trip, a camp, a graduation trip, a wandering trip, an employee trip, a school trip It can be mentioned that it is in the middle or on an excursion. Examples of TPO when the user is indoors include waking up, going to bed, returning to home, relaxing, injured, refreshing, eavesdropping, doing love, and doing household chores. , In a group, in a hustle, in a depressing state, in a busy situation, in a drowsiness, sick in sleep, lethargic, exercising, learning, stretching, I'm reading, I'm crazy about something, I'm playing an instrument, I'm cooking, I'm doing laundry, I'm listening to music, I'm operating a computer or smartphone, I'm playing a game Yes, watching TV, talking on the phone, talking, making a change, taking a bath, visiting, banqueting, having a meal, excited, lively , High tension, depressed, unwell, full of energy, in good shape, at work, raising children, playing hide-and-seek, playing tag, etc. In addition, as an example of TPO when the user is in a facility such as an aquarium, a zoo, a museum, an art gallery, or an observation facility, who the user is with (one, with children, with adults, with men, Such things as where to be in the facility with the woman, with the family, etc., what the user is doing (what they are interested in, what they are looking at, etc.). Examples of TPO when the user is in the shop window are: the user is approaching the shop window, who the user is with, what they are going to buy (bargain, recommended today) Etc.) can be mentioned. The examples of these TPOs may also be included in the attribute information about the user.

<Function block configuration>
FIG. 2 is a diagram showing an example of a functional block configuration of the IoT device 10 according to the present embodiment. The IoT device 10 according to the present embodiment includes a communication unit 101, a recommendation control unit 102, and a UI (User Interface) unit 103. The communication unit 101, the recommendation control unit 102, and the UI unit 103 are realized by the processor executing a program stored in the memory of the IoT device 10. The program can be stored in a non-transitory computer-readable storage medium.

  The IoT device 10 may be configured by one information processing device or may be configured by a plurality of information processing devices. In the case of being configured with a plurality of information processing devices, the functions performed by these functional blocks may be realized by the plurality of information processing devices operating in cooperation with each other.

  The communication unit 101 has a function of mutually communicating with the user terminal 20 via the communication network N. The recommendation control unit 102 has a function of performing various recommendations based on attribute information (first attribute information) managed by the IoT device 10. Further, the recommendation control unit 102 further includes an attribute information management unit 1021 and a natural language processing unit 1022. The attribute information management unit 1021 has a function of managing attribute information that the IoT device 10 should manage.

  The natural language processing unit 1022 has a function of recognizing a natural language. More specifically, the natural language processing unit 1022 uses, as a morpheme, a sentence using natural language input to the IoT device 10 (or a sentence generated by recognizing a voice input to the IoT device 10 by voice recognition). It includes a function of analyzing, a function of understanding an intention by interpreting a sentence decomposed into morphemes, a function of generating a response sentence (or a voice response sentence) using a natural sentence based on the understood intention. The generation of the response sentence can be realized, for example, by preparing a large number of conversation scenarios in advance and calling the conversation scenario corresponding to the input sentence. The natural language processing unit 1022 is used when the user terminal 20 and the IoT device 10 communicate with each other using a natural language instead of a dedicated communication protocol, or when a direct conversation is made between the IoT device 10 and the user. It is a functional unit used for the above and can be omitted in the present embodiment.

  The recommendation control unit 102 has a machine learning function, and has learned in advance, for example, information on hot-selling products that sell at this time, information on popular recipes, and the like. Also, the recommendation control unit 202 makes a predetermined recommendation by comparing the attribute information managed by the attribute information management unit 1021 with the information learned in advance. Any algorithm may be used for the machine learning, but for example, deep learning may be used.

  Here, specific examples of various recommendations performed by the recommendation control unit 102 will be described. It is assumed that the IoT device 10 is a smart window installed on a window glass of a clothing store, and the user inquires of the IoT device 10 about a recommended item for clothing. The recommendation control unit 102 compares the stock status of each product managed by the attribute information management unit 1021 with the ranking of the best-selling products in the current season, so that the highest ranking among the products in stock is obtained. Performs operations such as recommending certain hot-selling products.

  The UI unit 103 has a function of displaying various kinds of information on a screen included in the IoT device 10, a function of outputting a sound from a speaker, and a function of receiving a user's input from a touch panel and a microphone included in the IoT device 10.

  FIG. 3 is a diagram showing an example of a functional block configuration of the user terminal 20 according to the present embodiment. The user terminal 20 according to this embodiment includes a communication unit 201, a recommendation control unit 202, and a UI unit 203. The communication unit 201, the recommendation control unit 202, and the UI unit 203 are realized by the processor executing a program stored in the memory of the user terminal 20. The program can be stored in a non-transitory computer-readable storage medium.

  The communication unit 201, the recommendation control unit 202, and the UI unit 203 may all be provided in the user terminal 20, or a part of these functional blocks (for example, the recommendation control unit 202) may be provided. It may be configured to be provided on the external server side that can communicate with the user terminal 20. In the former case, the user terminal 20 may be called a "second device", and in the latter case, the user terminal 20 and the server may be collectively called a "second device".

  The communication unit 201 has a function of mutually communicating with the IoT device 10 via the communication network N.

  The recommendation control unit 202 has a function of recommending the user based on the attribute information (second attribute information) about the user. Further, the recommendation control unit 202 further includes an attribute information management unit 2021 and a natural language processing unit 2022. The attribute information management unit 2021 manages attribute information of a user who uses the user terminal 20. The attribute information management unit 2021 has a function of performing machine learning, and the attributes regarding the user's taste may be learned using the user's schedule information, action history, and the like.

  Here, a specific example of the learning content performed by the attribute information management unit 2021 will be described. For example, the attribute information management unit 2021 may learn the user's preference for the product by using the purchase history of the product purchased by the user in the past. For example, when the user has purchased the most red clothes in the past, the user's preference regarding clothes can be learned as “prefer red clothes”. Further, as another example, the attribute information management unit 2021 may learn the user's meal preference by combining the history of the position information of the user and the schedule information. For example, if a user often goes to a set restaurant on weekdays (at work) and often goes to an Italian restaurant on holidays, the user likes the set restaurant on weekdays and the Italian restaurant on holidays. You can learn.

  The attribute information management unit 2021 may allow the user to input the basic attributes of the user (gender, age, etc.) and the attributes indicating the behavior of the user, or may obtain them from an external server or the like. Alternatively, the information may be acquired using the GPS or the like included in the user terminal 20. In addition, the attribute information management unit 2021 acquires attributes (posture, calorie consumption, activity amount, sleep condition, heart rate, blood pressure, pulse rate, etc.) related to the biometric information of the user from the wearable device worn by the user. You may do it.

  The natural language processing unit 2022 has a function of recognizing a natural language. More specifically, the natural language processing unit 2022 uses, as a morpheme, a sentence using natural language input to the user terminal 20 (or a sentence generated by recognizing a voice input to the user terminal 20 by voice). It includes a function of analyzing, a function of understanding an intention by interpreting a sentence decomposed into morphemes, a function of generating a response sentence (or a voice response sentence) using a natural sentence based on the understood intention. The generation of the response sentence can be realized, for example, by preparing a large number of conversation scenarios in advance and calling the conversation scenario corresponding to the input sentence. The natural language processing unit 2022 is used when the user terminal 20 and the IoT device 10 communicate using natural language instead of a dedicated communication protocol, or when the user terminal 20 and the user directly have a conversation. It is a functional unit used for the above and can be omitted in the present embodiment.

  The recommendation control unit 202 has a machine learning function. For example, information that a user having such an attribute likes such a thing, or that a user having such an attribute should recommend such a thing. Information is learned in advance. The recommendation control unit 202 makes a recommendation to the user by comparing the user's attribute information managed by the attribute information management unit 2021 with previously learned information. Any algorithm may be used for the machine learning, but for example, deep learning may be used.

  Here, specific examples of various recommendations performed by the recommendation control unit 202 will be described. For example, it is assumed that the user inputs a request to buy clothes into the user terminal 20. The recommendation control unit 202 compares the attributes (sex, age) of the user managed by the attribute information management unit 2021 with the classification of each clothing attribute of other users who has been learned in advance, An operation is performed in which clothes that have the same attribute as the user and are preferred by another user are extracted and the clothes are recommended to the user.

  The UI unit 203 has a function of displaying various types of information on a screen included in the user terminal 20, a function of outputting sound from a speaker, and a function of receiving a user input from a touch panel, a microphone, or the like included in the user terminal 20.

  In the present embodiment, the recommendation control unit 102 of the IoT device 10 and the recommendation control unit 202 of the user terminal 20 cooperate to enable more appropriate recommendation to the user. Specifically, the recommendation control unit 202 of the user terminal 20 responds to the recommendation information notified by the recommendation control unit 102 of the IoT device 10 based on the attribute information of the user (for example, the notified recommendation information is The IoT device 10 is notified of information that the user does not suit the taste of the user and wants to propose other recommendation information, or information specifically indicating the taste of the user). Subsequently, the recommendation control unit 102 of the IoT device 10 notifies the user terminal 20 of the recommendation information including the content of the recommendation different from the recommendation information notified to the user terminal 20 last time, based on the notified response information. The IoT device 10 and the user terminal 20 perform a predetermined recommendation to the user by repeating such processing one or more times.

  For example, as described above, when the user inputs to the user terminal 20 a request to buy clothes, the recommendation control unit 202 of the user terminal 20 selects clothing that has the same attribute as the user and is preferred by another user. By extracting, an operation of recommending clothes assumed to be preferred by the user will be performed. On the other hand, when the user inquires the IoT device 10 about recommended products for clothing, the recommendation control unit 102 of the IoT device 10 performs an operation of recommending a hot-selling product that is the highest ranking among the products in stock. To do.

  That is, when the recommendation is performed by the user terminal 20 alone, it is possible to make a recommendation based on the attribute information of the user who uses the user terminal 20, but the recommendation is based on whether or not the actually recommended product can be purchased. Can't do. On the other hand, when the IoT device 10 is solely recommended, it is possible to recommend the most popular product among the products in stock, but it is not possible to perform the recommendation based on the attribute information of the user. However, when the IoT device 10 and the user terminal 20 cooperate with each other, it is possible to recommend, for example, a product that is assumed to be preferred by the user and that is a popular product among the products in stock. Become.

  As described above, in the present embodiment, the IoT device 10 and the user terminal 20 cooperate with each other, thereby making it possible to perform more appropriate recommendation according to the TPO or the like of the user while being personalized by the user.

<Operation outline and operation example>
Subsequently, an outline of an operation performed by the recommendation system according to the present embodiment and a specific operation example will be described. First, the outline of the operation will be described.

(Operation overview)
FIG. 4 is a sequence diagram showing an example of an operation outline performed by the recommendation system according to the present embodiment.

  First, communication is established between the communication unit 101 of the IoT device 10 and the communication unit 201 of the user terminal 20 (S101). For example, the communication unit 201 of the user terminal 20 searches for the IoT device 10 existing near the current position of the user terminal 20, and when the IoT device 10 existing nearby is detected, the detected IoT device 10 and the detected IoT device 10 are detected. Communication may be established between them. When a plurality of IoT devices 10 are searched, the communication unit 201 of the user terminal 20 may prompt the user to select the IoT device 10 that establishes communication.

  The IoT device 10 existing near the current position of the user terminal 20 may be detected by any method. For example, the communication unit 201 of the IoT device 10 transmits a beacon, and the user terminal 20 uses the beacon. You may make it search for the IoT device 10 which exists near oneself by searching for. Alternatively, a database indicating a place (latitude, longitude, etc.) where the IoT device 10 is installed is prepared in advance in an external server or the like, and the user terminal 20 accesses the database to determine a predetermined position from its own current position. The IoT device 10 existing in the range (for example, within a radius of 30 m) may be searched.

  In addition, the communication unit 201 of the user terminal 20 selects the IoT device 10 to establish communication from among the plurality of IoT devices 10 based on the user attribute information managed by the attribute information management unit 2021. Good. For example, when the schedule of going to the home appliance mass store A is registered in the scheduler of the user, the communication unit 201 of the user terminal 20 communicates with the IoT device 10 installed in the home appliance mass store A. It may be established. In addition, when the schedule of going out in the car is registered in the scheduler of the user, the communication unit 201 of the user terminal 20 includes the IoT device 10 installed in the car (for example, installed in the window of the driver's seat). However, communication may be established with the IoT device 10 that displays the recommended day trip destination on the driver's seat window. The communication unit 201 of the user terminal 20 may transmit an identifier (user ID or the like) that uniquely identifies a user to the IoT device 10 when establishing communication. The IoT device 10 can uniquely identify the recommended user.

  Subsequently, when the UI unit 203 of the user terminal 20 receives an instruction from the user that the recommendation information is desired, the recommendation control unit 202 transmits a message requesting the user to recommend to the IoT device 10. Yes (S102). After establishing the communication in step S101, the IoT device 10 sends to the user terminal 20 information indicating a category in which the IoT device 10 can recommend (for example, it is possible to recommend wines handled at the liquor store A). You may make it notify. Further, the UI unit 203 of the user terminal 20 may display the information indicating the notified category that the IoT device 10 can recommend on the display or the like so that the user can refer to the information.

  Further, when there are a plurality of categories that the IoT device 10 can recommend, the UI unit 203 of the user terminal 20 accepts the category for which the recommendation is desired from the user, and the recommendation control unit 202 requests the accepted category. It may be included in a message and transmitted to the IoT device 10.

  Then, the recommendation control unit 102 of the IoT device 10 performs the recommendation process based on the attribute information managed by the IoT device 10 and transmits the recommendation information including the content of the recommendation proposed to the user to the user terminal 20 ( S103). Whether or not the recommended control unit 202 of the user terminal 20 may accept the received recommended information by comparing the content of the recommendation included in the recommended information with the attribute information of the user (that is, the received recommended information is Whether to notify the user as it is) is determined. For example, when the content of the recommendation matches the attribute information of the user, the recommendation control unit 202 determines to accept the recommendation information (that is, notifies the user of the recommendation information as it is), and all or part of the content of the recommendation. Does not match the attribute information of the user (for example, the color of the recommended product does not match the color the user likes), the recommendation information cannot be accepted (that is, the recommendation information cannot be notified to the user as it is). May not be possible). When it is determined that the recommendation information is accepted, the UI unit 203 of the user terminal 20 displays the recommendation information notified from the IoT device 10 on a display or the like, and the process proceeds to step S107.

  If it is determined in step S103 that the recommendation information cannot be accepted, the recommendation control unit 202 determines the attribute information of the user (for example, the user's favorite color) about the portion of the content of the recommendation that does not match the attribute information of the user. ) Is transmitted to the IoT device 10 (S104).

  Subsequently, the recommendation control unit 102 of the IoT device 10 performs the recommendation process again based on the attribute information managed by the IoT device 10 and the user attribute information included in the response information, and obtains the recommendation information including the content of the recommendation. It transmits to the user terminal 20 (S105). The recommendation control unit 202 of the user terminal 20 compares the content of the recommendation included in the recommendation information with the attribute information of the user to determine again whether or not the recommendation information may be accepted. When the recommendation control unit 202 determines to accept the recommendation information, the recommendation information notified from the IoT device 10 is displayed on the display or the like on the UI unit 203 of the user terminal 20, and the process proceeds to step S107. On the other hand, when it is determined that the recommendation information cannot be accepted, the recommendation control unit 202 of the user terminal 20 performs the processing procedure of step S104 again. After that, the recommendation control unit 202 of the user terminal 20 determines to accept the recommendation information, or the recommendation control unit 102 of the IoT device 10 sends another recommendation information due to the reason that the information that can be recommended has been exhausted. The processing procedure of steps S104 and S105 is repeated until the user terminal 20 is notified of the information indicating that transmission is not possible.

  Then, the UI unit 203 of the user terminal 20 displays the recommendation information notified from the IoT device 10 on a display or the like (S106). Here, the recommendation control unit 202 may acquire the user's reaction (for example, liked or disliked the recommended content) via the UI unit 203. If the reaction of the user is negative, the recommendation control unit 202 of the user terminal 20 may learn that the recommendation information notified from the IoT device 10 has a negative reaction of the user. ..

  Subsequently, the recommendation control unit 202 of the user terminal 20 may transmit the acquired user reaction to the IoT device 10 (S107). The recommendation control unit 102 of the IoT device 10 may learn the reaction shown by the specific user who is using the user terminal 20 to the recommended content. For example, when a negative reaction is received from the user, by learning the negative reaction of the user, when the request for recommendation is received from the same user in the future, the recommendation information other than the negative reaction is received. It becomes possible to notify the recommendation information of. Alternatively, the recommendation control unit 102 of the IoT device 10 may learn by considering the received reaction of the specific user as the reaction shown by the unspecified user with respect to the recommended content. For example, when a negative reaction is received from a user, learning is performed by an unspecified user as a negative reaction, so that a negative reaction is received when a recommendation request is received from any user in the future. It becomes possible to notify the recommendation information other than the received recommendation information. The processing procedure of step S107 may be omitted.

  The recommendation control unit 202 of the user terminal 20 notifies the IoT device 10 that it has determined in step S103 or step S105 that the recommended information proposed by the IoT device 20 is accepted (S108), and the recommendation of the IoT device 10 is notified. The control unit 102 may display the recommendation information on a display or the like included in the IoT device 10 (S109). That is, the final recommendation information (recommendation information accepted by the recommendation control unit 202 of the user terminal 20) may be displayed on the screen of the user terminal 20 or the screen of the IoT device 10, or either one of them. It may be displayed only on the screen.

  In the processing procedure of steps S102 to S107 described above, the recommendation control unit 202 of the user terminal 20 and the recommendation control unit 102 of the IoT device 10 transmit and receive the recommendation information and the response information using a text message in natural language. Alternatively, the recommendation information and the response information may be transmitted and received using a dedicated communication protocol used for transmitting and receiving the recommendation information and the response information.

(Operation example)
Subsequently, a specific operation example performed by the recommendation system according to the present embodiment will be described. For the sake of convenience, the following operation example will be described on the assumption that the recommendation information and the response information are transmitted and received between the IoT device 10 and the user terminal 20 using a text message in natural language. Further, the description of the procedure for establishing communication between the IoT device 10 and the user terminal 20 is omitted.

[Operation example 1]
FIG. 5 is a diagram for explaining the first operation example. The first operation example is that the user A goes to the clothing store A to buy a coat, and the product name and image of the coat recommended for the user A are displayed on the window glass installed at the store of the clothing store A. It is supposed to be displayed.

  First, the recommendation control unit 202 of the user terminal 20 transmits a message indicating that the user A wants to know the recommended coat to the IoT device 10 based on the instruction of the user A (S201). Subsequently, the recommendation control unit 102 of the IoT device 10 determines a product to be recommended to the user A. Here, it is assumed that the attribute information managed by the IoT device 10 is “the number of stocks and the number of sales for each product”. The recommendation control unit 102 is based on a predetermined algorithm (here, it is assumed that this is an algorithm that selects the product that is in stock and has the highest number of sales), and the plurality of products handled by the clothing store A are selected. , Recommends the best-selling coat in stock at clothing store A. Next, the recommendation control unit 102 recommends that "Coat A (red duffle coat of 20,000 yen)," which is the best-selling coat in stock at the clothing store A, be recommended. The message shown is transmitted to the user terminal 20 (S202).

  Subsequently, the recommendation control unit 202 of the user terminal 20 refers to the attribute information of the user A to recognize that the preference of the clothing of the user A is “blue”, and the user A wears a blue coat. A message indicating the desire is transmitted to the IoT device 10 (S203). Then, the recommendation control unit 102 of the IoT device 10 puts "Coat B (blue duffel coat of 20,000 yen)" in stock at the clothing store A and the best-selling blue coat, A message indicating that the product is recommended as a recommended product is transmitted to the user terminal 20 (S204).

  Subsequently, the recommendation control unit 202 of the user terminal 20 refers to the attribute information of the user A to grasp that the preference of the clothing of the user A is “trench coat”, and the user A wears the trench coat. A message indicating the desire is transmitted to the IoT device 10 (S205). Then, the recommendation control unit 102 of the IoT device 10 puts “Coat C (blue trench coat of 40,000 yen)” that is in stock at the clothing store A and is the best-selling blue trench coat. A message indicating that the product is recommended as a recommended product is transmitted to the user terminal 20 (S206).

  Subsequently, the recommendation control unit 202 of the user terminal 20 refers to the attribute information of the user A to grasp that the clothing that the user A has purchased up to now is 30,000 yen or less, and the user A has 30,000. A message indicating that a coat less than a circle is desired is transmitted to the IoT device 10 (S207). Subsequently, the recommendation control unit 102 of the IoT device 10 is in stock at the clothing store A and is the best-selling blue trench coat of 30,000 yen or less. A message indicating that “blue trench coat)” is recommended as a recommended product is transmitted to the user terminal 20 (S208).

  Then, the recommendation control unit 202 of the user terminal 20 determines that the recommended information indicating that the court D is recommended is accepted because the recommended court D matches the attribute information of the user, and the recommendation information is accepted. Is transmitted to the IoT device 10 (S209).

  Subsequently, the UI unit 103 of the IoT device 10 indicates the recommended product for the user A, for example, the product name, the image, and the price of the coat D, along with a message such as "I recommend this trench coat to customers". It is displayed on the window glass installed in the store of the store A (S210).

  According to the operation example described above, the user terminal 20 and the IoT device 10 cooperate with each other, so that the hot-selling products in the clothing store A are not simply recommended but are personalized based on the preference of the user A. The product will be recommended.

  In addition, in the operation example described above, the recommendation control unit 202 of the user terminal 20 may collectively notify the preference of the user A. That is, the recommendation control unit 202 may notify the IoT device 10 with one message that the user A desires a blue trench coat whose price is 30,000 yen or less.

[Operation example 2]
FIG. 6 is a diagram for explaining the second operation example. The second operation example is a situation in which the user A returns home and requests the refrigerator to display a recommended recipe using ingredients managed by the refrigerator (IoT device 10) installed in the home. Is assumed.

  First, the recommendation control unit 202 of the user terminal 20 transmits a message indicating that the user A wants to know the recipe to the IoT device 10 based on the instruction of the user A (S301). Subsequently, the recommendation control unit 102 of the IoT device 10 determines a recipe to be recommended to the user A. Here, it is assumed that the attribute information managed by the IoT device 10 is “the foodstuff in the refrigerator and the amount of each foodstuff”. Based on a predetermined algorithm, the recommendation control unit 102 determines a plurality of recipes that can be made from the ingredients in the refrigerator, and selects any of the determined plurality of recipes, “cook A (tomato pasta)”. A message indicating that “Recipe” is recommended as a recommended recipe is transmitted to the user terminal 20 (S302).

  Subsequently, the recommendation control unit 202 of the user terminal 20 refers to the attribute information of the user A to recognize that the user A does not like tomatoes, and the user A desires to cook a dish that does not use tomatoes. The message shown is transmitted to the IoT device 10 (S303). Next, the recommendation control unit 102 of the IoT device 10 recommends a “cooking B (carbonara) recipe” that is a recipe that does not use tomato, among a plurality of recipes that can be made from the ingredients in the refrigerator. A message indicating that the recipe is recommended is transmitted to the user terminal 20 (S304).

  Subsequently, the recommendation control unit 202 of the user terminal 20 refers to the attribute information of the user A, grasps that the user A has high blood pressure, and issues a message indicating that he/she wants a dish with low calories, to the IoT device 10. (S305). Subsequently, the recommendation control unit 102 of the IoT device 10 is a recipe that does not use tomato and has a lower calorie than the food B (carbonara) among a plurality of recipes that can be made from the food materials in the refrigerator. A message indicating that “Cook C (oden) recipe” is recommended as a recommended recipe is transmitted to the user terminal 20 (S306). Note that the recommendation control unit 202 and the recommendation control unit 102 hold a list of calories for each recipe in advance, and determine whether the food has a high calorie or a low calorie depending on whether the calorie is equal to or more than a predetermined value. It may be determined whether or not.

  Subsequently, the recommendation control unit 202 of the user terminal 20 determines to accept the recommendation information that the recipe of the dish C is recommended as the recommended recipe, and transmits a message indicating that the recommendation information is accepted to the IoT device 10 (S307). ).

  Then, the UI unit 203 of the user terminal 20 displays, as a recommended recipe for the user A, for example, the name of the dish C and a specific recipe on the screen of the user terminal 20 together with a message such as “Today's recommended recipe”. It is displayed (S308).

  According to the specific example described above, the cooperation of the user terminal 20 and the IoT device 10 does not simply recommend the recipe of the dish that can be cooked with the ingredients in the refrigerator, but rather favors the preference of the user A. Based on that, personalized recipes will be recommended.

  The embodiments described above are for facilitating the understanding of the present invention and are not for limiting the interpretation of the present invention. Each element included in the embodiment and its arrangement, material, condition, shape, size and the like are not limited to the exemplified ones and can be appropriately changed. Further, the configurations shown in different embodiments can be partially replaced or combined.

  In the present embodiment, the “unit” does not simply mean a physical means, but also includes a case where the function of the “unit” is realized by software. Even if the function of one "section" or device is realized by two or more physical means or devices, the functions of two or more "sections" or devices are realized by one physical means or device. May be.

10... IoT device, 20... User terminal, 101... Communication part, 102... Recommendation control part, 103... UI part, 201... Communication part, 202... Recommendation control part, 203... UI part, 1021... Attribute information management part, 1022 ...Natural language processing unit, 2021...attribute information management unit, 2022...natural language processing unit

Claims (5)

A recommendation system including a first device used by an unspecified user and a second device used by a specific user,
The first device is
A first recommendation control unit configured to perform recommendation based on first attribute information managed by the first device, and notify the second device of recommendation information including the content of the recommendation,
The second device is
A second recommendation control unit configured to notify the first device of response information based on the recommendation information notified from the first device and second attribute information about the specific user,
When the second recommendation control unit determines whether or not to recommend the recommendation information to the specific user based on the recommendation information and the second attribute information, and determines that the recommendation information should not be recommended. Notifying the specific user of an attribute to be recommended as the response information,
The first recommendation control unit notifies the second device of other recommendation information including recommendation content different from the recommendation information based on the response information,
Recommendation system. The first attribute information is information indicating an attribute for each of a plurality of products handled by the first device,
The first recommendation control unit notifies the recommendation information indicating the attribute of any one of the plurality of products, which is determined based on a predetermined algorithm.
The second recommendation control unit determines whether or not to recommend the product indicated by the recommendation information to the specific user, based on the attribute of the product indicated by the recommendation information and the second attribute information. When it is determined that it should not be recommended, the attribute of the product to be recommended is notified to the specific user as the response information,
The recommendation system according to claim 1. The second device is a communication unit that communicates with the first device, and based on the second attribute information, a communication unit that selects the first device that establishes communication from a plurality of the first devices, Has,
The recommendation system according to claim 1. The first recommendation control unit notifies the recommendation information to the second device using natural language,
The second recommendation control unit notifies the response information to the first device using natural language,
The recommendation system according to any one of claims 1 to 3. A recommendation method executed by a recommendation system including a first device used by an unspecified user and a second device used by a specific user, comprising:
A step in which the first device makes a recommendation based on first attribute information managed by the first device, and notifies the second device of recommendation information including the content of the recommendation;
The step of the second device notifying the first device of response information based on the recommendation information notified from the first device and second attribute information related to the specific user , wherein the recommendation information and the recommendation information Based on the second attribute information, it is determined whether or not the recommendation information should be recommended to the specific user, and when it is determined that the recommendation information should not be recommended, the attribute to be recommended to the specific user is determined. Notifying as the response information ,
A step in which the first device notifies the second device of other recommendation information including content of a recommendation different from the recommendation information based on the response information;
Recommending method having.