Evaluation value estimation system, evaluation value estimation method, evaluation value estimation device, evaluation value estimation program, and storage medium recording the program.

Abstract

This estimation system S comprises an emotion recognition unit 12 that recognizes an emotional valence intensity and arousal level of a user toward an object, and an evaluation value estimation unit 16 that estimates a user evaluation value using a learning model that receives input of the emotional valence intensity and arousal level of the user and outputs a user evaluation value.

Description

Technical Field

[0001] The present invention relates to an evaluation value estimation system, an evaluation value estimation method, an evaluation value estimation device, an evaluation value estimation program, and a storage medium storing the program for estimating the user's motivation for a target object.

Background Art

[0002] Generally, even for the same object, the content and intensity of the emotion vary from person to person. Also, even for one's own emotions, it is difficult to accurately and quantitatively express such emotions. Furthermore, since the emotion towards a certain object changes over time, it is desirable to express the emotion promptly. On the other hand, there is a desire to quantitatively grasp the emotions of a large number of people towards a certain object in a short time.

[0003] To meet such a desire, various methods for measuring emotions have been proposed. For example, in the technique described in Patent Document 1, a model having joints imitating the human body is used to grasp the emotion of the user based on how the user bends which joints.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, when a model is used to express a user's emotions, as in the technology described in Patent Document 1, a certain explanation of the method of expression is required. Furthermore, even if the expression by the model is the same, the emotions associated with that expression may differ from user to user depending on cultural background and other factors. Moreover, expressing a user's emotions through a model requires a certain amount of time to manipulate the model, and there is a risk that the emotions may change from those felt when interacting with the object.

[0006] Therefore, data on emotions collected using such models risked being less generalizable. For example, if an analysis was performed based on data collected from users living in a certain region, the results of that analysis may not necessarily be applicable to users living in other regions.

[0007] Furthermore, using such complex models presented a problem: the user input process became cumbersome, making it difficult to collect data from a large number of users.

[0008] The present invention has been made in view of the above points, and aims to provide an evaluation value estimation system, evaluation value estimation method, evaluation value estimation device, evaluation value estimation program, and a storage medium on which the program is recorded, which can quickly and easily collect data on user emotions in a highly versatile manner, and furthermore, can easily analyze the collected data. [Means for solving the problem]

[0009] The evaluation value estimation system of the present invention is A system for estimating user evaluation values, which are values ​​representing a user's motivation towards an object, An emotion recognition unit that recognizes the emotional valence intensity and arousal level of the user toward the object, The system includes an evaluation value estimation unit that estimates the user evaluation value using a learning model that takes the emotional valence intensity and arousal level of the user as input and outputs the user evaluation value. , basic configurationLet's assume that.

[0010] Here, "valence" is a bipolar concept that indicates a qualitative difference in the emotion evoked (for example, a positive value such as attraction (pleasure) or a negative value such as aversion (displeasure) towards an object). Furthermore, "valence intensity" is the strength to which that emotion is felt (Haranobe, Yoshihiro (2015). Measurement of general valence concept intensity using paired comparison method: A preliminary study using unpleasant emotions. Tezukayama Gakuin University Faculty of Human Sciences Research Annual Report, (17)).

[0011] Furthermore, "arousal" here refers to a concept that has two extremes: excitement (high arousal) and calmness (low arousal), and indicates the intensity of the emotions evoked (Bradley, MM, Greenwald, MK, Petry, MC, & Lang, PJ (1992). Remembering Pictures: Pleasure and Arousal in Memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 379-390.).

[0012] Furthermore, "evaluation value" here refers to a value representing the desire for the subject or similar items. As evaluation values, for example, past consumption frequency, likelihood of recommending to others, willingness to consume, willingness to purchase, etc., can be expressed as clear numerical values ​​using a point scale with arbitrarily set values.

[0013] Thus, the evaluation value estimation system of the present invention employs universal parameters such as emotional valence and arousal level. Therefore, the user evaluation values ​​output by inputting these parameters are highly versatile values ​​that are not dependent on cultural background or other factors. Furthermore, emotional valence and arousal level are parameters that can be easily obtained without the use of special equipment.

[0014] Furthermore, since these parameters are simplified compared to those used to express complex emotions, using them for analysis makes the analysis process easier and more quantitative compared to using conventional parameters that express emotions (for example, user-generated text).

[0015] Therefore, this system allows for the rapid and easy collection of user emotional data in a highly versatile manner, and furthermore, the collected data can be easily analyzed. This is particularly effective when the evaluation subject involves complex factors such as cultural background in addition to individual physical condition and preferences, such as cooking.

[0016] Furthermore, in the evaluation value estimation system of the present invention, The emotion recognition unit includes an emotion model presentation unit that presents to the user an emotion model, which is a planar or three-dimensional model including at least a first axis indicating emotional valence intensity and a second axis indicating arousal level, in a format in which any coordinate in the emotion model can be selected, and a selected coordinate recognition unit that recognizes the selected coordinate, which is the coordinate selected by the user from the emotion model when the object is presented to the user. The evaluation value estimation unit preferably takes the selected coordinates as input.

[0017] In addition to the basic configuration mentioned above , the evaluation value estimation system of the present invention Mu is , The aforementioned user evaluation value includes values ​​for multiple evaluation items, The learning model takes one of the aforementioned emotional valence levels and arousal level as input and outputs values ​​for the aforementioned multiple evaluation items. Characterized by .

[0018] In addition, since the parameter of coordinates is a parameter that can be intuitively understood by the user, there is no need to give any special explanation to the user about the expression method when collecting data. As a result, parameters related to emotions that are likely to change can be quickly collected before such changes occur. Also, since the parameter of coordinates is simple and clear, when analysis is performed using such coordinates as parameters, the analysis process can be carried out more easily.

[0019] In addition, as the third axis in the case where the emotion model is a three-dimensional model, for example, an axis indicating the degree of spread of arousal may be adopted (Toshio Sato, Kazuya Horige (2018). Examination of an emotion scale using a three-dimensional model. The 82nd Annual Convention of the Japanese Psychological Association, 721).

[0020] Also, in the evaluation value estimation system of the present invention, the user evaluation value includes values for a plurality of evaluation items, it is preferable that the learning model takes one of the emotional value intensity and the arousal degree as an input and outputs values for the plurality of evaluation items.

[0021] Here, the "evaluation item" is, for example, an item based on the five senses of a human for the target thing. For example, when the target thing is food, the taste of the food (taste), aroma / flavor (smell), appearance (vision), texture (touch), sound when eating or cooking (hearing), etc. become evaluation items.

[0022] The user evaluation value does not necessarily have to be a value for one evaluation item, and it may include values for a plurality of evaluation items in this way. And when a plurality of evaluation items are adopted for the user evaluation value in this way, a multi-faceted evaluation becomes possible, so the analysis result using the user evaluation value can be made more accurate and objective.

[0023] Also, in the evaluation value estimation system of the present invention, when the evaluation value includes values for a plurality of evaluation items, Preferably, the system includes an attractiveness value estimation unit that, by inputting one of the aforementioned emotional valence intensity and arousal level, weights each of the values ​​for the plurality of evaluation items output from the learning model, and estimates an attractiveness value by summing each of the weighted values.

[0024] In addition to the basic configuration mentioned above , the evaluation value estimation system of the present invention Mu is , The aforementioned user evaluation value includes values ​​for multiple evaluation criteria, The learning model takes one of the aforementioned emotional valence levels and the aforementioned arousal level as input and outputs a value based on the aforementioned multiple evaluation criteria. Characterized by .

[0025] Furthermore, in the evaluation value estimation system of the present invention, if the user evaluation value includes values ​​for multiple evaluation items, The aforementioned object is food, Preferably, the aforementioned multiple evaluation items include at least one of the following: an evaluation item based on taste, an evaluation item based on smell, an evaluation item based on sight, and an evaluation item based on touch.

[0026] Generally, food evaluations involve multiple factors such as taste (gustatory), aroma / flavor (olfactory), appearance (visual), texture (tactile), and sounds made when eating or cooking (auditory), making the resulting evaluations complex. Therefore, a system that allows for easy and instantaneous collection of emotional data as input data, and outputs user evaluation values ​​that are easy to handle during analysis, can facilitate and accurately perform such food evaluations.

[0027] Furthermore, in the evaluation value estimation system of the present invention, The aforementioned user evaluation value includes values ​​for multiple evaluation criteria, Preferably, the learning model takes one of the aforementioned emotional valence levels and the aforementioned arousal level as input and outputs a value based on the plurality of evaluation criteria.

[0028] Here, "evaluation criteria" refers to the standards for evaluating the willingness to purchase or similar items. Therefore, the "evaluation value" in this case is the value obtained from those evaluation criteria. Examples of evaluation criteria may include past consumption frequency, likelihood of recommending to others, willingness to consume, and willingness to purchase.

[0029] User evaluation values ​​do not necessarily have to be a single value; they may include values ​​for multiple evaluation criteria (i.e., multiple values). By adopting values ​​for multiple evaluation criteria in this way, it becomes possible to evaluate a single item from multiple perspectives, thus making the analysis results using those user evaluation values ​​more objective and accurate.

[0030] Furthermore, in the evaluation value estimation system of the present invention, if the user evaluation value includes values ​​for multiple evaluation criteria, Preferably, the system includes an attractiveness value estimation unit that, by inputting one of the aforementioned emotional valence intensity and arousal level, weights each of the values ​​for the plurality of evaluation criteria output from the learning model, and recognizes the sum of each of the weighted values ​​as an attractiveness value.

[0031] Depending on the object for which user evaluation values ​​are being estimated and the purpose for which the user evaluation values ​​will be used, the evaluation criteria that should be emphasized may differ. Therefore, when outputting an attractiveness value, which is a single value obtained by summing up user evaluation values, applying different weights to each evaluation criterion according to the purpose of use before calculating the attractiveness value can make the output attractiveness value, and consequently the analysis results using that attractiveness value, more accurate.

[0032] Furthermore, in the evaluation value estimation system of the present invention, if the evaluation value includes values ​​for multiple evaluation criteria, The values ​​for the aforementioned multiple evaluation criteria may include at least one of the following: values ​​for past consumption frequency, values ​​for the likelihood of recommending to others, values ​​for the likelihood of consumption, and values ​​for the likelihood of purchase.

[0033] Furthermore, in the evaluation value estimation system of the present invention, when an emotion model is adopted as an interface for inputting emotions, which is a planar model or a three-dimensional model that includes at least a first axis indicating emotional valence intensity and a second axis indicating arousal level, The learning model may also be correlation data that shows the correlation between reference coordinates, which are coordinates selected from the emotion model, and reference evaluation values, which are values ​​regarding willingness towards the target object or similar objects having the same or similar attributes as the target object.

[0034] Furthermore, the evaluation value estimation device of the present invention is An evaluation value estimation device that estimates a user evaluation value, which is a value representing the user's motivation towards an object, An emotion recognition unit that recognizes the emotional valence intensity and arousal level of the user toward the object, The system includes an evaluation value estimation unit that estimates the user evaluation value using a learning model that takes the user's emotional valence intensity and arousal level as input and outputs the user evaluation value, The aforementioned user evaluation value includes values ​​for multiple evaluation items, The learning model takes one of the aforementioned emotional valence levels and arousal levels as input and outputs values ​​for the aforementioned multiple evaluation items. It is characterized by the following. Furthermore, the evaluation value estimation device of the present invention is An evaluation value estimation device that estimates a user evaluation value, which is a value representing the user's motivation towards an object, An emotion recognition unit that recognizes the emotional valence intensity and arousal level of the user toward the object, The system includes an evaluation value estimation unit that estimates the user evaluation value using a learning model that takes the user's emotional valence intensity and arousal level as input and outputs the user evaluation value, The aforementioned user evaluation value includes values ​​for multiple evaluation criteria, The learning model is characterized by taking one of the aforementioned emotional valence levels and the aforementioned arousal level as input and outputting a value based on the aforementioned multiple evaluation criteria.

[0035] Furthermore, the evaluation value estimation method of the present invention is A method for estimating user evaluation values, which are values ​​representing a user's motivation towards an object, The emotion recognition unit performs the steps of recognizing the emotional valence intensity and arousal level of the user toward the object, The evaluation value estimation unit includes the step of estimating the user evaluation value using a learning model that takes the emotional valence intensity and arousal level provided by the user as input and outputs the user evaluation value. 、 The aforementioned user evaluation value includes values ​​for multiple evaluation items, The learning model takes one of the aforementioned emotional valence levels and arousal level as input and outputs values ​​for the aforementioned multiple evaluation items. It is characterized by the following. Furthermore, the evaluation value estimation method of the present invention is A method for estimating user evaluation values, which are values ​​representing a user's motivation towards an object, The emotion recognition unit performs the steps of recognizing the emotional valence intensity and arousal level of the user toward the object, The evaluation value estimation unit includes the step of estimating the user evaluation value using a learning model that takes the emotional valence intensity and arousal level provided by the user as input and outputs the user evaluation value, The aforementioned user evaluation value includes values ​​for multiple evaluation criteria, The learning model is characterized by taking one of the aforementioned emotional valence levels and the aforementioned arousal level as input and outputting a value based on the aforementioned multiple evaluation criteria.

[0036] Furthermore, the evaluation value estimation program of the present invention is In an evaluation value estimation program that causes a computer to execute an evaluation value estimation method for estimating a user evaluation value, which is a value representing the user's willingness to deal with an object, To the aforementioned computer, The emotion recognition unit performs the steps of recognizing the emotional valence intensity and arousal level of the user toward the object, The evaluation value estimation unit takes the emotional valence intensity and arousal level provided by the user as input and uses a learning model that outputs the user evaluation value to estimate the user evaluation value. Let's execute it 、 The aforementioned user evaluation value includes values ​​for multiple evaluation items, The learning model takes one of the aforementioned emotional valence levels and arousal levels as input and outputs values ​​for the aforementioned multiple evaluation items. It is characterized by the following. Furthermore, the evaluation value estimation program of the present invention is In an evaluation value estimation program that causes a computer to execute an evaluation value estimation method for estimating a user evaluation value, which is a value representing the user's willingness to deal with an object, To the aforementioned computer, The emotion recognition unit performs the steps of recognizing the emotional valence intensity and arousal level of the user toward the object, The evaluation value estimation unit performs the step of estimating the user evaluation value using a learning model that takes the emotional valence intensity and arousal level provided by the user as input and outputs the user evaluation value. The aforementioned user evaluation value includes values ​​for multiple evaluation criteria, The learning model is characterized by taking one of the aforementioned emotional valence levels and the aforementioned arousal level as input and outputting a value based on the aforementioned multiple evaluation criteria.

[0037] Furthermore, the evaluation value estimation device of the present invention is An evaluation value estimation device that estimates a user evaluation value, which is a value representing the user's motivation towards an object, An emotion recognition unit that recognizes the emotional valence intensity and arousal level of the user toward the object, The system is characterized by comprising: an evaluation value estimation unit that estimates the user evaluation value using a learning model that takes the emotional valence intensity and arousal level of the user as input and outputs the user evaluation value; and a learning model that estimates the user evaluation value.

[0038] Furthermore, the evaluation value estimation method of the present invention is A method for estimating user evaluation values, which are values ​​representing a user's motivation towards an object, The emotion recognition unit performs the steps of recognizing the emotional valence intensity and arousal level of the user toward the object, The evaluation value estimation unit is characterized by comprising the step of estimating the user evaluation value using a learning model that takes the emotional valence intensity and arousal level provided by the user as input and outputs the user evaluation value.

[0039] Furthermore, the evaluation value estimation program of the present invention is In an evaluation value estimation program that causes a computer to execute an evaluation value estimation method for estimating a user evaluation value, which is a value representing the user's willingness to deal with an object, To the aforementioned computer, The emotion recognition unit performs the steps of recognizing the emotional valence intensity and arousal level of the user toward the object, The evaluation value estimation unit is characterized by performing the step of estimating the user evaluation value using a learning model that takes the emotional valence intensity and arousal level provided by the user as input and outputs the user evaluation value.

[0040] Furthermore, the recording medium of the present invention is A recording medium characterized by recording the above-mentioned evaluation value estimation program and allowing the computer to read the evaluation value estimation program. [Brief explanation of the drawing]

[0041] [Figure 1] An explanatory diagram showing the schematic configuration of the estimation system according to the embodiment. [Figure 2] A block diagram showing the configuration of the processing unit of the estimation system in Figure 1. [Figure 3] Figure 1 is an illustrative diagram showing an example of an image displayed on the user terminal when the estimation system recognizes user information. [Figure 4] Figure 1 is an illustrative image showing an example of an image displayed on the user's terminal when the estimation system recognizes the user's emotions. [Figure 5A] A schematic diagram showing an example of an emotion model used to recognize emotions in the estimation system shown in Figure 1. [Figure 5B] A schematic diagram showing an example of an emotion model used to recognize emotions in the estimation system related to the modified example. [Figure 6] Figure 1 is an illustrative image showing an example of an image displayed on the user terminal when the estimation system recognizes a reference evaluation value. [Figure 7] Figure 1 is an illustrative diagram showing an example of correlation data recognized by the estimation system. [Figure 8] A flowchart showing the process that the estimation system performs when collecting correlation data (Figure 1). [Figure 9] A flowchart showing the processes performed by the estimation system in Figure 1 when estimating attractiveness. [Figure 10] Figure 1 is an illustrative image showing an example of an image displayed on the client terminal after the estimation system has estimated the attractiveness value. [Figure 11] Sample images used in the data collection experiment to generate the learning model for the estimation system shown in Figure 1. [Figure 12] The experimental data used to generate the learning model for the estimation system shown in Figure 1 is the data related to purchase probability score. [Figure 13] The experimental data used to generate the learning model for the estimation system shown in Figure 1 is data related to willingness to take home. [Figure 14]The experimental data used to generate the learning model for the estimation system shown in Figure 1 includes data related to the degree of recommendation to others (Net Promoter Score). [Modes for carrying out the invention]

[0042] The evaluation value estimation system according to the embodiment (hereinafter referred to as "Estimation System S") and the evaluation value estimation method implemented using it will be described below with reference to the drawings.

[0043] In this embodiment, we will describe a case in which the estimation system S estimates the attractiveness value of an object (a dish in this embodiment) for a user U, and uses that attractiveness value in a service that presents it to a client engaged in marketing or other activities.

[0044] In the following explanation, "evaluation value" refers to a value representing the willingness towards the target object or similar objects described later. Specifically, it may include values ​​that are clearly expressed using point scales with arbitrarily set values, such as past consumption frequency (e.g., Frequency of past consumption), likelihood of recommending to others (e.g., Net promoter score), willingness to consume (e.g., Willingness to take home), and willingness to purchase (e.g., Purchase Probability Score, Willingness to buy, the Juster scale). The estimation system S employs these values.

[0045] Furthermore, the "reference evaluation value" is an evaluation value used to generate the correlation data (learning model) described later, and represents a value indicating user U's motivation towards the target object or similar objects as a test user.

[0046] In contrast, the "user evaluation value" is an evaluation value estimated by the evaluation value estimation unit 16, which will be described later, and represents a value regarding the user U's willingness to the object of interest as the person whose evaluation is estimated by the estimation system S (i.e., the target of marketing).

[0047] Furthermore, "target object" here refers to the object being evaluated (i.e., the object for which the user evaluation value is being estimated). "Similar object" refers to an object that has the same or similar attributes as the target object.

[0048] In this embodiment, the object of study is a dish. Therefore, similar objects include dishes that have the same or similar taste as the dish in which the object of study is the object of study, flavors that have the same or similar aroma and flavor as the dish, photographs of the dish or a similar dish, dishes that have the same or similar texture as the dish, and dishes that make the same or similar sounds when eaten or prepared. Furthermore, the attributes of the object of study or similar objects in this embodiment may include taste, aroma and flavor, appearance, and texture.

[0049] Incidentally, evaluations of food (and by extension, food products) are generally based on multiple criteria, such as taste (gustatory), aroma / flavor (olfactory), appearance (visual), texture (tactile), and sounds made when eating or cooking (auditory), and the resulting evaluations tend to be complex. However, even for things that are prone to such complex evaluations, the evaluation value estimation system of the present invention can be used to perform evaluations easily and accurately.

[0050] However, the evaluation value estimation system of the present invention is not limited to such a configuration, and the estimated evaluation values ​​can not necessarily be used only for estimating attractiveness values ​​based on user evaluation values ​​of dishes.

[0051] Therefore, for example, the evaluation value estimation system of the present invention may estimate user evaluation values ​​from a large number of users for a new food product or a non-food product (e.g., clothing), and refer to those user evaluation values ​​themselves for marketing purposes of the new product.

[0052] [System Outline] The schematic configuration of the estimation system S will be described below with reference to Figures 1 and 2.

[0053] As shown in Figure 1, the estimation system S is a computer system and is composed of a server 1 owned by the provider of the service provided by the estimation system S.

[0054] Server 1 is configured to communicate with user terminals 2, such as smartphones and tablets used by user U, and client terminals 3, such as personal computers used by clients, via the internet, public telephone lines, etc.

[0055] Furthermore, the evaluation value estimation system of the present invention is not limited to being composed of a single server; it is sufficient if any of the terminals constituting the evaluation value estimation system are configured to be equipped with the processing unit described later.

[0056] Therefore, for example, the evaluation value estimation system may be configured using multiple servers. Alternatively, for example, at least one processing unit or at least a part of the functions of the processing unit may be implemented in at least one of the user terminals and client terminals (the client terminal), and the evaluation value estimation system may be configured in cooperation with the server, or with at least one of the user terminals and client terminals alone.

[0057] Furthermore, for example, functions corresponding to the user-side input unit 20 and user-side output unit 21 provided in the user terminal 2 in this embodiment (see Figure 2) may be provided in a terminal having a processing unit, and configured as an independent evaluation value estimation device. In addition, functions corresponding to the client terminal-side input unit (not shown) and client-side output unit 30 provided in the client terminal 3 (see Figure 2) may be provided in that evaluation value estimation device.

[0058] As shown in Figure 2, the user terminal 2, which can communicate with the estimated system S, server 1, includes a user-side input unit 20 and a user-side output unit 21 as functions (processing units) realized by at least one of the implemented hardware configuration and program.

[0059] In this embodiment, the user-side input unit 20 and user-side output unit 21 of the user terminal 2 are assumed to be touch panels (see Figure 3, etc.).

[0060] However, the user terminal is not limited to such a configuration; it can be any device capable of receiving information input from the user and outputting information for presentation to the user. For example, the user terminal may be configured to enable input and output using a touch panel, keyboard, microphone, camera, speaker, etc.

[0061] Furthermore, the client terminal 3, which can communicate with the server 1, which is the estimated system S, includes a client-side input unit (not shown) and a client-side output unit 30 as functions (processing units) realized by at least one of the implemented hardware configuration and program.

[0062] In this embodiment, the client-side input section of the client terminal 3 is assumed to be a keyboard and touchpad, and the client-side output section 30 is assumed to be a monitor (see Figure 10).

[0063] However, the client terminal is not limited to such a configuration; it can be any device capable of outputting information to the client. For example, the client terminal may be configured to output information using a speaker or similar device.

[0064] Furthermore, user terminals and client terminals are not limited to those composed of physical objects in the real world; they may also be composed of physical objects in a virtual world, or they may be composed of a combination of physical and virtual objects. When configured using physical objects in a virtual world, input and output occur in the virtual world, or in a space that combines the virtual and real worlds.

[0065] [Configuration of each processing unit] Next, the processing units that constitute the estimation system S will be described with reference to Figures 2 to 7.

[0066] As shown in Figure 2, Server 1 includes, as functions (processing units) realized by at least one of the implemented hardware configuration and program, an attribute recognition unit 10, an object presentation unit 11, an emotion recognition unit 12, a reference evaluation value recognition unit 13, a correlation data generation unit 14, a correlation data storage unit 15, an evaluation value estimation unit 16, and an attractiveness value estimation unit 17.

[0067] As shown in Figure 3, the attribute recognition unit 10 presents user U with a questionnaire in an answerable format via the user terminal 2 to confirm user U's attributes, and recognizes user U's attributes based on the answers to the questionnaire.

[0068] Here, a user's "attributes" refer to information that directly or indirectly influences their evaluation of things. Examples include age, gender, whether or not they have food allergies, physical information such as height and weight, and place of origin (i.e., cultural background). The estimation system S uses age, gender, place of origin, and food allergies as its attributes.

[0069] The attribute recognition unit 10 may be configured to recognize user U's attributes by referring to information previously obtained about user U, without referring to the questionnaire and its responses. Here, the previously obtained information may be, for example, information entered by user U in other systems provided by the provider of the service by the estimation system S.

[0070] The object presentation unit 11 presents information about the object whose evaluation value is being recognized by user U via the user terminal 2. Specifically, in this embodiment, the object is a dish, so the object presentation unit 11 presents an image of the dish, as well as information about places where the dish can be eaten.

[0071] In the estimation system S, information about the target object is presented in this manner. However, in the present invention, the method of presenting the target object to the user may be changed as appropriate, as long as it is a method that allows the user to recognize the target object.

[0072] Therefore, for example, as will be described later in this embodiment, the test user may be presented with the object itself (more specifically, the dish itself), while the user attempting to estimate the attractiveness value may be presented with only an image of the object.

[0073] However, in the evaluation value estimation system of the present invention, it is preferable to present the same information to both the test user and the user who intends to estimate the attractiveness value, and further, it is preferable to present the information in the same environment (for example, whether or not to present the object itself). Also, in the evaluation value estimation system of the present invention, if the object itself is not to be presented on the system (for example, if only the object itself is presented without presenting the information), the object presentation unit may be omitted.

[0074] The emotion recognition unit 12 recognizes the emotions of user U when using the estimation system S. The emotion recognition unit 12 includes an emotion model presentation unit 12a and a coordinate recognition unit 12b (selected coordinate recognition unit).

[0075] As shown in Figure 4, the emotion model presentation unit 12a presents an emotion model, which is a planar model including a first axis indicating emotional valence intensity and a second axis indicating arousal level, to the user U via the user terminal 2 in a format in which the user can select any coordinate in the emotion model. The estimation system S employs EmojiGrid®, shown in Figure 5A, as the emotion model to be presented.

[0076] Here, "valence" is a bipolar concept that indicates a qualitative difference in the emotion evoked (for example, a positive value such as attraction (pleasure) or a negative value such as aversion (displeasure) towards an object). Furthermore, "valence intensity" is the strength to which that emotion is felt (Haranobe, Yoshihiro (2015). Measurement of general valence concept intensity using paired comparison method: A preliminary study using unpleasant emotions. Tezukayama Gakuin University Faculty of Human Sciences Research Annual Report, (17)).

[0077] Furthermore, "arousal" here refers to a concept that has two extremes: excitement (high arousal) and calmness (low arousal), and indicates the intensity of the emotions evoked (Bradley, MM, Greenwald, MK, Petry, MC, & Lang, PJ (1992). Remembering Pictures: Pleasure and Arousal in Memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 379-390.).

[0078] Furthermore, the emotion model in this invention is not limited to EmojiGrid, but may be any planar or three-dimensional model including a first axis representing emotional valence intensity and a second axis representing arousal level.

[0079] Therefore, for example, Russell's Circle of Emotion model may be adopted as the emotion model, as shown in the modified version in Figure 5B. Furthermore, a model that adapts the Circle of Emotion model, in which colors or emojis are used instead of text, may also be adopted.

[0080] Furthermore, in this embodiment and its modifications, a planar model was used as the emotion model. However, the emotion model of the present invention is not limited to such a configuration, and may also be a three-dimensional model of such planar models.

[0081] Furthermore, when the emotion model is made into a three-dimensional model, a third axis may be adopted, for example, an axis indicating the degree of arousal (Sato, Toshio & Horike, Kazuya (2018). Examination of an emotion scale using a three-dimensional model. The 82nd Annual Meeting of the Japanese Psychological Association, 721).

[0082] The coordinate recognition unit 12b recognizes the coordinates (selected coordinates or reference coordinates) in the emotion model selected by user U via user terminal 2. In estimation system S, the coordinate recognition unit 12b recognizes the coordinates specified by user U touching the emotion model displayed on the touch panel, which is the user-side input unit 20 and user-side output unit 21 of user terminal 2.

[0083] The coordinate recognition unit 12b then recognizes the recognized coordinates themselves as the emotions of user U at that time. In other words, the coordinate recognition unit 12b does not specifically identify what emotions are associated with the coordinates.

[0084] Thus, in the estimation system S, the emotion recognition unit 12 employs an emotion model that is a planar model including at least a first axis indicating emotional valence intensity and a second axis indicating arousal level, as an interface for user U to input emotions.

[0085] As a result, in this estimation system S, user U can easily and accurately express their emotions as clear values ​​in the form of coordinates. Consequently, data regarding user U's emotions can be easily and accurately collected.

[0086] However, the emotion recognition unit in the evaluation value estimation system of the present invention is not limited to such a configuration, and only needs to recognize the emotional valence intensity and arousal level of the user towards the object.

[0087] Therefore, for example, the emotion recognition unit may recognize at least one of the emotional valence intensity and arousal level as a numerical value. Furthermore, the method of recognition does not necessarily have to be based on user input; it may be recognized by estimation or calculation from the user's facial expressions, actions, etc.

[0088] The reference evaluation value recognition unit 13 recognizes reference evaluation values ​​for the objects presented when user U uses the estimation system S.

[0089] Specifically, as shown in Figure 6, the reference evaluation value recognition unit 13 presents a questionnaire, provided for each predetermined evaluation criterion, to the user U via the user terminal 2 in an answerable format, and recognizes the answer as a reference evaluation value.

[0090] Here, "evaluation criteria" refers to the criteria for evaluating the willingness to purchase or buy a similar item. Therefore, the "evaluation value" in this case is the value obtained using those evaluation criteria. Possible evaluation criteria include, for example, frequency of past consumption, likelihood of recommending to others (for example, Net promoter score), willingness to purchase (for example, Willingness to take home), and willingness to buy (for example, Purchase Probability Score, Willingness to buy, the Juster scale).

[0091] The estimation system S is configured to ask user U for at least one of the following values ​​for each of the predetermined evaluation items (in this embodiment, taste, aroma / flavor, and appearance): a value for past consumption frequency, a value for recommendation to others, a value for consumption intent, and a value for purchase intent.

[0092] Here, "evaluation items" refer to items based on the five human senses in relation to the object in question. For example, if the object in question is food, the evaluation items would include the taste (gustatory), aroma / flavor (olfactory), appearance (visual), texture (tactile), and sounds heard when eating or cooking the food (auditory).

[0093] The correlation data generation unit 14 generates correlation data (learning model) based on the emotions (reference coordinates) of user U as a test user recognized by the emotion recognition unit 12 and the reference evaluation values ​​of user U recognized by the reference evaluation value recognition unit 13.

[0094] In the estimation system S, as shown in Figure 7, correlation data is generated as two-dimensional data showing the correlation between coordinates selected from the emotion model and specific values ​​of the evaluation values. This correlation data is generated for each user attribute, for each of the predetermined evaluation items (in this embodiment, taste, aroma / flavor, and appearance), and for each of the multiple evaluation criteria.

[0095] The correlation data generated in this way takes the selected coordinates chosen from the emotion model by user U, who is the user whose attractiveness value is estimated, as input, and outputs a user evaluation value.

[0096] Furthermore, the learning model in the evaluation value estimation system of the present invention is not limited to such correlation data, but can take emotional valence intensity and arousal level as input and output user evaluation values.

[0097] Therefore, for example, the learning model may be a predictive model that is generated by machine learning using coordinates selected from an emotion model and reference evaluation values, which are values ​​representing the willingness towards the target object or similar object, as training data, and outputs evaluation values ​​from the input coordinates. Furthermore, the input is not limited to coordinates, but may also be the values ​​of emotional valence intensity and arousal level themselves.

[0098] Furthermore, although not adopted in estimation system S, as described later, when estimating user evaluation values ​​for target objects as the target of marketing, the generated learning model may be updated using the user evaluation values ​​estimated in that process.

[0099] Furthermore, in the correlation data (learning model) of this embodiment, for a single input data (coordinates in this embodiment, and consequently, emotional valence intensity and arousal level), values ​​for multiple evaluation criteria are output for each of the multiple evaluation items. However, the learning model in an evaluation value estimation system is not limited to this configuration; it is sufficient if it takes emotional valence intensity and arousal level as input and outputs user evaluation values.

[0100] Therefore, for example, a learning model may output one value for one input data, or it may output one value for multiple input data, or it may output multiple values ​​for multiple input data.

[0101] The correlation data storage unit 15 stores the correlation data generated by the correlation data generation unit 14, classifying it according to the attributes of user U. Specifically, the correlation data storage unit 15 associates the correlation data with the attributes of user U, who input the reference coordinates and reference evaluation values ​​when generating the correlation data, and stores them accordingly.

[0102] The evaluation value estimation unit 16 estimates the user evaluation value of the object by user U based on the coordinates selected from the emotion model and the correlation data obtained from the correlation data storage unit 15.

[0103] In the estimation system S, the evaluation value estimation unit 16 estimates, as user evaluation values, values ​​for each of the predetermined evaluation items (in this embodiment, taste, aroma / flavor, and appearance), including values ​​for past consumption frequency, recommendation to others, willingness to consume, and willingness to purchase.

[0104] This is to enable a multifaceted evaluation by recognizing values ​​for multiple evaluation criteria for each of the multiple evaluation items related to a single object, thereby making the analysis results using those user evaluation values ​​highly objective and accurate.

[0105] However, the evaluation value estimation system of the present invention is not limited to such configurations. For example, the evaluation value estimation system may recognize only one evaluation item, or it may recognize only a value for one evaluation criterion.

[0106] Incidentally, the evaluation value estimation unit 16 acquires and recognizes correlation data (learning model) corresponding to the user U's attributes from the correlation data storage unit 15, and uses that correlation data to estimate the user evaluation value. This is because the accuracy of estimating the user evaluation value is further improved when the user's attributes are the same or similar.

[0107] However, the evaluation value estimation system of the present invention is not limited to such a configuration, and any system that estimates user evaluation values ​​using a learning model is acceptable. For example, the evaluation value estimation system may be configured to use the same learning model even if the users are different, without changing the learning model according to the user's attributes when estimating user evaluation values.

[0108] The attractiveness value estimation unit 17 estimates the attractiveness value based on the user evaluation value estimated by the evaluation value estimation unit 16.

[0109] Specifically, the attractiveness value estimation unit 17 first applies weights to each of the multiple evaluation criteria (such as values ​​for past consumption frequency) included in each of the values ​​for multiple evaluation items (in this embodiment, taste, aroma / flavor, appearance, and texture) estimated by the evaluation value estimation unit 16, based on predetermined rules according to the user U's attributes (for example, different weights for each evaluation criterion). The unit then estimates the attractiveness value for each of the multiple evaluation items for the target object (in this embodiment, food) by summing up each of these weighted values.

[0110] Subsequently, the attractiveness value estimation unit 17 further weights the attractiveness value for each of the multiple evaluation items estimated in this manner based on predetermined rules according to the attributes of user U (for example, different weights for each evaluation item), and estimates the attractiveness value of the object itself by summing up each of these weighted values.

[0111] This is because the evaluation criteria or evaluation items that should be emphasized may differ depending on the object being estimated for user evaluation and the purpose for which the user evaluation is used.

[0112] Therefore, when outputting an attractiveness value, which is a single value obtained by summing up user evaluation values, configuring the system to calculate the attractiveness value after weighting it according to the purpose of use, etc., will make the output attractiveness value, and consequently the analysis results using that attractiveness value, more accurate. The same applies when using a predictive model instead of correlational data as the learning model.

[0113] However, the evaluation value estimation system of the present invention is not limited to such a configuration, and the method for calculating the attractiveness value may be set as appropriate.

[0114] Therefore, for example, weighting may be performed not by user attributes as in this embodiment, but according to the type of object, or by rules based on other criteria, or by combining multiple rules. Alternatively, the attractiveness value may be simply the sum of user evaluation values ​​without weighting. Furthermore, for example, weighting may be performed only for predetermined evaluation criteria, or only for predetermined evaluation items.

[0115] Alternatively, instead of summing up multiple user ratings, one user rating that should be emphasized based on the type of object may be determined and used as the attractiveness rating. Furthermore, the attractiveness rating may not be estimated, and one or more user ratings may be presented directly. In such cases, the attractiveness rating estimation unit may be omitted.

[0116] Regarding the weighting rules, for example, one could investigate which evaluation items are given more weight based on at least one of the user's attributes and the type of object being studied, and then establish rules based on the results of that investigation. Alternatively, one could investigate which evaluation criteria are given more weight based on at least one of the user's attributes or the type of object being studied, and then establish rules based on the results of that investigation. Furthermore, combinations of these rules are also possible.

[0117] [Processes executed in each processing unit when generating a learning model] Next, referring to Figures 2-4, 6, and 8, we will explain the processes performed in each processing unit when the estimation system S generates correlation data, which is the learning model.

[0118] In this process, User U is a test user who pre-evaluates the target object in order to collect data for generating correlation data, which is a learning model.

[0119] As explained below, the estimation system S presents the target object itself to the test user, User U, and performs the process of generating correlation data, which is the learning model. However, it is also possible to present similar objects instead of the target object and perform the same process. Furthermore, if there are multiple test users, it is possible to present the target object to some of the test users and similar objects to others and perform the same process.

[0120] In this process, first, the attribute recognition unit 10 presents user U with a questionnaire in an answerable format via the user terminal 2, as shown in Figure 3, to confirm user U's attributes (Figure 8 / STEP100).

[0121] Next, the attribute recognition unit 10 recognizes the attributes of user U based on the answers to the questionnaire (Figure 8 / STEP 101).

[0122] Next, the object presentation unit 11 presents information about the object, which is the dish (image, name, ingredients, etc.), to the user U via the user terminal 2 (Figure 8 / STEP 102).

[0123] In this process according to this embodiment, information about the food, which is the object of the process, is presented, and the food itself is provided to user U.

[0124] Next, the emotion model presentation unit 12a of the emotion recognition unit 12 presents the emotion model, which is a planar model including a first axis indicating emotional valence intensity and a second axis indicating arousal level, to the user U via the user terminal 2 in a format in which the user can select any coordinate in the emotion model (Figure 8 / STEP 103).

[0125] Next, the coordinate recognition unit 12b of the emotion recognition unit 12 recognizes the coordinates selected by the user U from the emotion model via the user terminal 2 as reference coordinates (Figure 8 / STEP 104).

[0126] Specifically, as shown in Figure 4, first, the emotion model presentation unit 12a displays the emotion model (EmojiGrid) on the touch panel of the user terminal 2. Then, when user U touches any point on this emotion model, the coordinate recognition unit 12b determines that the coordinates of that point have been selected and recognizes those coordinates as reference coordinates.

[0127] Next, the reference evaluation value recognition unit 13 presents user U with a questionnaire regarding the reference evaluation value provided by user U, via the user terminal 2, in a format that user U can answer (Figure 8 / STEP 105).

[0128] Next, the reference evaluation value recognition unit 13 recognizes the reference evaluation value based on the answers to the questionnaire (Figure 8 / STEP 106).

[0129] Specifically, as shown in Figure 6, first, the reference evaluation value recognition unit 13 displays several questions on the touch panel of the user terminal 2 to ask about values ​​such as past consumption frequency. Then, when user U selects one of the answers to a question, the reference evaluation value recognition unit 13 recognizes the selected answer as the reference evaluation value in the evaluation criteria corresponding to that question.

[0130] In the estimation system S, the questionnaire conducted by the reference evaluation value recognition unit 13 is performed sequentially for each of the predetermined evaluation items (in this embodiment, taste, aroma / flavor, appearance, and texture).

[0131] Next, the correlation data generation unit 14 generates correlation data (learning model) based on the emotion of user U recognized by the emotion recognition unit 12 (reference coordinates) and the reference evaluation value of user U recognized by the reference evaluation value recognition unit 13 (Figure 8 / STEP 107).

[0132] In the estimation system S, this correlation data is generated for each of the predetermined evaluation items (in this embodiment, taste, aroma / flavor, appearance, and texture) in a number of ways equal to the number of types of reference evaluation values ​​(i.e., evaluation criteria).

[0133] Next, the correlation data generation unit 14 classifies the generated correlation data according to the attributes recognized by the attribute recognition unit 10, stores it in the correlation data storage unit 15 (Figure 8 / STEP 108), and ends this process.

[0134] In this embodiment, information about the target object, a dish (image, name, ingredients, etc.), is presented to user U, who is a test user, and the above processing is performed when user U actually eats the dish.

[0135] However, the process for generating the learning model in the evaluation value estimation system of the present invention does not necessarily have to be performed at such a timing; it can be performed before the execution of the process for estimating the user evaluation value described later. For example, the process for generating the model may be performed not only when the target object or similar object is presented, but also when it is acquired (for example, when it is purchased).

[0136] Furthermore, the process of recognizing user U's attributes (STEP 100 and STEP 101) can be performed at any time before the process of storing the correlation data (STEP 108). For example, it can be performed in advance independently of the series of processes, or it can be performed after the generation of the correlation data (after STEP 107).

[0137] Furthermore, in this embodiment, we have described a case where the learning model is correlation data showing the correlation between reference coordinates and reference evaluation values ​​by user U, who is a test user. However, after performing the same processing as in STEP 102 to STEP 106, machine learning may be performed using the reference coordinates obtained in that processing and the reference evaluation values ​​by user U, who is a test user, as training data to generate a prediction model, and that prediction model may be used as the learning model.

[0138] Furthermore, in this embodiment, the estimation system S generates a learning model by performing the processing described in STEP 100 to STEP 108 above. However, the evaluation value estimation system of the present invention is not limited to this configuration and may be configured to utilize a separately generated learning model.

[0139] [Processes executed in each processing unit when estimating user evaluation values ​​and attractiveness values] Next, referring to Figures 2-4, 9, and 10, we will explain the processes executed in each processing unit when the estimation system S estimates user evaluation values ​​and attractiveness values.

[0140] In this process, User U is the target of marketing for the target object. The estimated attractiveness value for User U is then presented to the client conducting the marketing via client terminal 3. In the following process, a similar object may be used instead of the target object, and the results of that process may be used to estimate at least one of the user evaluation value and attractiveness value of the target object.

[0141] In this process, first, the attribute recognition unit 10 presents user U with a questionnaire in an answerable format via the user terminal 2, as shown in Figure 3, to confirm user U's attributes (Figure 9 / STEP200).

[0142] Next, the attribute recognition unit 10 recognizes the attributes of user U based on the answers to the questionnaire (Figure 9 / STEP201).

[0143] Next, the object presentation unit 11 presents information about the object, which is the dish (image, name, ingredients, etc.), to the user U via the user terminal 2 (Figure 9 / STEP202).

[0144] In this process in this embodiment, only information about the food, which is the subject matter, is presented, and the food itself is not provided to user U. However, when generating the correlation data (learning model) used in this process, it is preferable to present it to user U, who is the target of marketing, in the same way as it was presented to the test user.

[0145] Next, the emotion model presentation unit 12a of the emotion recognition unit 12 presents the emotion model, which is a planar model including a first axis indicating emotional valence intensity and a second axis indicating arousal level, to the user U via the user terminal 2 in a format in which the user can select any coordinate in the emotion model (Figure 9 / STEP203).

[0146] Next, the coordinate recognition unit 12b of the emotion recognition unit 12 recognizes the coordinates selected by the user U from the emotion model via the user terminal 2 as selected coordinates (Figure 9 / STEP204).

[0147] Specifically, as shown in Figure 4, first, the emotion model presentation unit 12a displays the emotion model (EmojiGrid) on the touch panel of the user terminal 2. Then, when user U touches any point on this emotion model, the coordinate recognition unit 12b determines that the coordinates of that point have been selected and recognizes those coordinates as the selected coordinates.

[0148] Next, the evaluation value estimation unit 16 acquires and recognizes correlation data corresponding to the attributes recognized by the attribute recognition unit 10 from the correlation data storage unit 15 (Figure 9 / STEP205).

[0149] Specifically, the evaluation value estimation unit 16 selects and recognizes the correlation data from among the multiple correlation data stored in the correlation data storage unit 15 in which at least one attribute of the linked attribute matches or is similar to an attribute recognized by the attribute recognition unit 10, and which has the largest number of matching or similar items.

[0150] Next, the evaluation value estimation unit 16 recognizes the user evaluation value based on the recognized correlation data and the user U's emotions (selected coordinates) recognized by the emotion recognition unit 12 (Figure 9 / STEP206).

[0151] Specifically, the evaluation value estimation unit 16 takes the recognized selected coordinates as input and outputs values ​​for each of the multiple evaluation criteria for each of the predetermined evaluation items (in this embodiment, taste, aroma / flavor, appearance, and texture) from the recognized correlation data.

[0152] Next, the evaluation value estimation unit 16 assigns weights to each of the multiple user evaluation values ​​it has recognized, based on the attributes recognized by the attribute recognition unit 10 (Figure 9 / STEP207).

[0153] Specifically, for example, the evaluation value estimation unit 16 recognizes a predetermined weighting rule for each attribute and weights each of the values ​​in the multiple evaluation criteria included in each user evaluation value according to that rule.

[0154] Next, the attractiveness value estimation unit 17 calculates the attractiveness value based on the weighted user evaluation values ​​(Figure 9 / STEP208).

[0155] Specifically, the attractiveness value estimation unit 17 estimates the attractiveness value for each evaluation item of the target object by summing up the values ​​of multiple evaluation criteria included in the user evaluation value for each evaluation item.

[0156] Furthermore, the attractiveness value estimation unit 17 weights the attractiveness value for each of the multiple estimated evaluation items according to the weighting rules obtained in STEP 207, and then sums these values ​​to estimate the attractiveness value of the object itself.

[0157] Next, as shown in Figure 10, the attractiveness value estimation unit 17 presents the calculated attractiveness value to the client U via the client terminal 3 as the attractiveness value of the object (Figure 9 / STEP209), and then terminates the process.

[0158] As explained above, the estimation system S and the evaluation value estimation method performed using the estimation system S employ universal parameters such as emotional valence and arousal level. Therefore, the user evaluation values ​​output by inputting these parameters are highly generalizable values ​​that are not dependent on cultural background or other factors. Furthermore, emotional valence and arousal level are parameters that can be easily obtained without the use of special equipment.

[0159] Furthermore, since these parameters are simplified compared to those used to express complex emotions, using them for analysis makes the analysis process easier and more quantitative compared to using conventional parameters that express emotions (for example, user-generated text).

[0160] Therefore, this system and method allow for the rapid and easy collection of user U's emotional data in a highly versatile manner, and furthermore, the collected data can be easily analyzed.

[0161] In the evaluation system S, user evaluation values ​​are defined as values ​​for multiple evaluation criteria for each of the multiple evaluation items. Furthermore, the evaluation system S estimates all of these values.

[0162] However, the evaluation value estimation system of the present invention is not limited to such a configuration, and at least one of multiple user evaluation values ​​may be recognized (specifically estimated, acquired, etc.) using another system. Specifically, for example, if the object is a dish, the evaluation value related to taste may be estimated by the evaluation value estimation system, and the evaluation value related to appearance may be recognized by actual sales figures.

[0163] [Example of experiment] Next, referring to Figures 11 to 14, we will describe the experiment to collect actual data on evaluation criteria and the experimental data obtained from that experiment.

[0164] This experiment corresponds to steps 100 to 106 of the aforementioned [processing performed in each processing unit when generating a learning model], as explained with reference to Figures 7 and 8. Therefore, the experimental data obtained from this experiment will be used to create correlation data showing the correlation between reference coordinates and reference evaluation values ​​by the test user, User U (see Figure 8 / Steps 107 to 108), and as training data for generating a prediction model.

[0165] The subjects of this experiment were 156 men and women aged 18 or older who belonged to a specified domain (specifically, who were citizens of a specified country). The gender breakdown of the subjects was 54.5% male and 45.5% female. The mean age of the subjects was 40.2 ± SD (standard deviation) 9.5 years.

[0166] In this experiment, participants were presented with images of six different foods (i.e., similar items to the target dish) as shown in Figure 11. They were then asked to input their emotions (emotional valence and arousal level) using an EmojiGrid (see Figure 4) and complete a questionnaire regarding evaluation criteria (see Figure 6). Note that Figure 11 is an illustrative image; in the actual experiment, photographs of the food were used.

[0167] In other words, this experiment used "appearance" as an evaluation criterion. Furthermore, this experiment used Purchase Probability Score, Willingness to take home, and Net Promoter Score as evaluation criteria.

[0168] In the "Purchase Probability Score" shown in Figure 12, the question "Are you likely to purchase this food item?" was used as the questionnaire, and the evaluation score obtained from this questionnaire was an 11-point scale, with 0 points representing "Definitely not buying" and 10 points representing "Definitely buying".

[0169] Furthermore, in the "Willingness to take home" section shown in Figure 13, the questionnaire question "How much would you like to take this food home?" was used, and the evaluation score obtained from this questionnaire was an 11-point scale, with "Not at all" being 0 points and "Strongly agree" being 10 points.

[0170] Furthermore, for the "Net Promoter Score" shown in Figure 14, the questionnaire question used was, "To what extent would you recommend this food to close friends and family?" The evaluation value obtained from this questionnaire was an 11-point scale, with "Not at all agree" being 0 points and "Strongly agree" being 10 points.

[0171] Figures 12 to 14 show the collected experimental data. In each figure, the top left data plots the emotions entered (coordinates selected in the EmojiGrid) for all evaluation values. The top right data plots the emotions entered when the evaluation value is 0 and 1. The middle left data plots the emotions entered when the evaluation value is 2, 3 and 4. The middle right data plots the emotions entered when the evaluation value is 5 and 6. The bottom left data plots the emotions entered when the evaluation value is 7 and 8. The bottom right data plots the emotions entered when the evaluation value is 9 and 10.

[0172] As shown in Figures 12-14, a clear and significant correlation was found between evaluation values ​​and emotions. Therefore, it was thought that a meaningful learning model could be generated using this data.

[0173] As mentioned above, the estimation system S employs an emotion model (EmojiGrid), which is a planar model that includes at least a first axis indicating emotional valence intensity and a second axis indicating arousal level, as an interface for user U to input emotions. This allows user U to easily and accurately express their emotions as clear values ​​in the form of coordinates.

[0174] Furthermore, since the coordinate parameter is a parameter that user U can intuitively understand, there is no need to provide users with any special explanation about how to represent the data when collecting it. As a result, the estimation system S can quickly collect parameters related to emotions, which are prone to change, before those changes occur.

[0175] Furthermore, since coordinates are a simple and clear parameter, using them as parameters in an analysis can make the analysis process even easier.

[0176] [Other embodiments] Although the illustrated embodiments have been described above, the present invention is not limited to these embodiments.

[0177] For example, in the above embodiment, the case in which the estimation system S is a single computer system was described. However, the present invention also includes an evaluation value estimation program for causing any one or more computers to execute the aforementioned evaluation value estimation method, and a recording medium for recording the program and which allows the program to be read by a computer used by a user or the like.

[0178] Furthermore, in the above embodiment, the attractiveness value of the object of user U, which is the target of marketing, is presented to the client, who is engaged in marketing or the like. However, the evaluation value estimation system of the present invention is not limited to this configuration.

[0179] Therefore, for example, the system may be configured to present the user evaluation value itself, or both the user evaluation value and the attractiveness value, estimated using the evaluation value estimation system of the present invention, to the user themselves. Alternatively, at least one of the estimated user evaluation value and attractiveness value may be used directly as data for machine learning in marketing, etc., rather than being presented to anyone. [Explanation of Symbols]

[0180] 1...Server, 2...User terminal, 3...Client terminal, 10...Attribute recognition unit, 11...Target object presentation unit, 12...Emotion recognition unit, 12a...Emotion model presentation unit, 12b...Coordinate recognition unit (selected coordinate recognition unit), 13...Reference evaluation value recognition unit, 14...Correlation data generation unit, 15...Correlation data storage unit, 16...Evaluation value estimation unit, 17...Attractiveness value estimation unit, 20...User-side input unit, 21...User-side output unit, 30...Client-side output unit, S...Estimation system (evaluation value estimation system), U...User.

Claims

1. A system for estimating user evaluation values, which are values ​​representing a user's motivation towards an object, An emotion recognition unit that recognizes information on the user's emotional valence intensity and level of arousal towards the object, The system includes an evaluation value estimation unit that estimates the user evaluation value using a pre-trained estimation model that takes the user's emotional valence intensity and arousal level information as input and outputs the user evaluation value as an estimated value, The user evaluation value includes a value indicating the user's enthusiasm for the object, based on multiple evaluation items that are based on the five human senses for the object. The aforementioned pre-trained model for estimation is a model constructed by training a predetermined model with information on emotional valence intensity and arousal level toward the object and a value indicating motivation toward the object as training data, and is characterized in that, at the time of estimation, it takes as input one piece of information on emotional valence intensity and arousal level toward the object from the user, and outputs estimated values ​​for the multiple evaluation items.

2. In the evaluation value estimation system according to claim 1, An evaluation value estimation system characterized by comprising an attractiveness value estimation unit that inputs one piece of information on emotional valence intensity and arousal level, weights each of the values ​​for the plurality of evaluation items output from the learning model, and estimates an attractiveness value by summing each of the weighted values.

3. In the evaluation value estimation system according to claim 1, The aforementioned object is food, An evaluation value estimation system characterized in that the plurality of evaluation items include at least one of the following: an evaluation item based on taste, an evaluation item based on smell, an evaluation item based on sight, and an evaluation item based on touch.

4. A system for estimating user evaluation values, which are values ​​representing a user's motivation towards an object, An emotion recognition unit that recognizes information on the user's emotional valence intensity and level of arousal towards the object, The system includes an evaluation value estimation unit that estimates the user evaluation value using a pre-trained estimation model that takes the user's emotional valence intensity and arousal level information as input and outputs the user evaluation value, The user evaluation value includes a value indicating the degree of multiple evaluation criteria that show the type of motivation towards the object, with respect to evaluation items based on the five human senses regarding the object. The pre-trained model for estimation is a model constructed by training a predetermined model with information on emotional valence intensity and arousal level toward the object and a value indicating motivation toward the object as training data, and is characterized in that, at the time of estimation, it takes as input one piece of information on emotional valence intensity and arousal level toward the object from the user, and outputs as an estimated value a value indicating the degree of the multiple evaluation criteria relating to the evaluation item.

5. In the evaluation value estimation system according to claim 4, An evaluation value estimation system characterized by comprising an attractiveness value estimation unit that, by inputting one of the aforementioned emotional valence intensity and arousal level, weights each of the values ​​indicating the degree of the plurality of evaluation criteria output from the pre-trained estimation model, and estimates an attractiveness value by summing each of the weighted values.

6. In the evaluation value estimation system according to claim 4, An evaluation value estimation system characterized in that the value indicating the degree of the aforementioned multiple evaluation criteria includes at least one of the following: a value for past consumption frequency, a value for the likelihood of recommending to others, a value for the likelihood of consumption, and a value for the likelihood of purchase.

7. In the evaluation value estimation system according to claim 1 or claim 4, The emotion recognition unit includes an emotion model presentation unit that presents to the user an emotion model, which is a planar or three-dimensional model including at least a first axis indicating the emotional valence intensity and a second axis indicating the arousal level, in a format in which the user can select any coordinate in the emotion model, and a selected coordinate recognition unit that recognizes the selected coordinates representing the emotional valence intensity and the arousal level, which the user selected from the emotion model when the object is presented to the user. The evaluation value estimation system is characterized in that the evaluation value estimation unit takes the selected coordinates as input.

8. In the evaluation value estimation system according to claim 7, The evaluation value estimation system is characterized in that the pre-trained model for estimation is correlation data showing the correlation between reference coordinates, which are coordinates selected from the emotion model, and reference evaluation values, which are values ​​for willingness towards the target object or similar objects having the same or similar attributes as the target object.

9. In the evaluation value estimation system according to claim 8, The aforementioned reference coordinates are coordinates selected from the emotion model by a test user who has previously evaluated the subject object or similar object, for the subject object or similar object. The evaluation value estimation system is characterized in that the correlation data is data showing the correlation between the reference coordinates and the reference evaluation value by the test user.

10. In the evaluation value estimation system according to claim 7, The evaluation value estimation system is characterized in that the pre-trained estimation model is generated by machine learning using reference coordinates, which are coordinates selected from the emotion model, and reference evaluation values, which are values ​​regarding willingness towards the target object or similar objects having the same or similar attributes as the target object, as training data, and is a predictive model that outputs the user evaluation value from the input selected coordinates.

11. In the evaluation value estimation system according to claim 10, The aforementioned reference coordinates are coordinates selected from the emotion model by a test user who has previously evaluated the subject object or similar object, for the subject object or similar object. The evaluation value estimation system is characterized in that the prediction model is generated by machine learning using the reference coordinates and the reference evaluation values ​​provided by the test user as training data.

12. An evaluation value estimation device that estimates a user evaluation value, which is a value representing the user's motivation towards an object, An emotion recognition unit that recognizes information on the user's emotional valence intensity and level of arousal towards the object, The system includes an evaluation value estimation unit that estimates the user evaluation value using a pre-trained estimation model that takes the user's emotional valence intensity and arousal level information as input and outputs the user evaluation value as an estimated value, The user evaluation value includes a value indicating the user's enthusiasm for the object, based on multiple evaluation items that are based on the five human senses for the object. The aforementioned pre-trained model for estimation is a model constructed by training a predetermined model with information on emotional valence intensity and arousal level toward the object and a value indicating motivation toward the object as training data, and is characterized in that, at the time of estimation, it takes as input one piece of information on emotional valence intensity and arousal level toward the object from the user, and outputs estimated values ​​for the multiple evaluation items.

13. An evaluation value estimation device that estimates a user evaluation value, which is a value representing the user's motivation towards an object, An emotion recognition unit that recognizes information on the user's emotional valence intensity and level of arousal towards the object, The system includes an evaluation value estimation unit that estimates the user evaluation value using a pre-trained estimation model that takes the user's emotional valence intensity and arousal level information as input and outputs the user evaluation value, The user evaluation value includes a value indicating the degree of multiple evaluation criteria that show the type of motivation towards the object, with respect to evaluation items based on the five human senses regarding the object. The pre-trained model for estimation is a model constructed by training a predetermined model with information on emotional valence intensity and arousal level toward the object and a value indicating motivation toward the object as training data, and is characterized in that, at the time of estimation, it takes as input one piece of information on emotional valence intensity and arousal level toward the object from the user, and outputs as an estimated value a value indicating the degree of the multiple evaluation criteria relating to the evaluation item.

14. A method for estimating user evaluation values, which are values ​​representing a user's motivation towards an object, The emotion recognition unit recognizes information regarding the emotional valence intensity and arousal level of the user towards the object, The evaluation value estimation unit includes the step of estimating the user evaluation value using a pre-trained estimation model that takes the user's emotional valence intensity and arousal level information as input and outputs the user evaluation value as an estimated value. The user evaluation value includes a value indicating the user's enthusiasm for the object, based on multiple evaluation items that are based on the five human senses for the object. The pre-trained model for estimation is a model constructed by training a predetermined model with information on emotional valence intensity and arousal level toward the object and a value indicating motivation toward the object as training data, and is characterized in that, at the time of estimation, it takes as input one piece of information on emotional valence intensity and arousal level toward the object from the user and outputs the values ​​for the multiple evaluation items as estimated values.

15. A method for estimating user evaluation values, which are values ​​representing a user's motivation towards an object, The emotion recognition unit recognizes information regarding the emotional valence intensity and arousal level of the user towards the object, The evaluation value estimation unit includes the step of estimating the user evaluation value using a pre-trained estimation model that takes the user's information on emotional valence intensity and arousal level as input and outputs the user evaluation value, The user evaluation value includes a value indicating the degree of multiple evaluation criteria that show the type of motivation towards the object, with respect to evaluation items based on the five human senses regarding the object. The pre-trained model for estimation is a model constructed by training a predetermined model with information on emotional valence intensity and arousal level toward the object and a value indicating motivation toward the object as training data, and is characterized in that, at the time of estimation, it takes as input one piece of information on emotional valence intensity and arousal level toward the object from the user, and outputs as an estimated value a value indicating the degree of the multiple evaluation criteria relating to the evaluation item.

16. An evaluation value estimation program that causes a computer to execute an evaluation value estimation method for estimating a user evaluation value, which is a value representing the user's willingness to deal with an object, When the evaluation estimation program is loaded by the computer, an emotion recognition unit and an evaluation estimation unit are constructed within the computer. The emotion recognition unit performs the step of recognizing information on the user's emotional valence intensity and arousal level towards the object, The evaluation value estimation unit includes the step of estimating the user evaluation value using a pre-trained estimation model that takes the user's emotional valence intensity and arousal level information as input and outputs the user evaluation value as an estimated value, The user evaluation value includes a value indicating the user's willingness towards the object for a plurality of evaluation items based on the five human senses for the object, and the pre-trained model for estimation is a model constructed by training a predetermined model with information on emotional valence intensity and arousal level for the object and a value indicating willingness towards the object as training data, and is characterized in that, at the time of estimation, it takes one piece of information on emotional valence intensity and arousal level for the object from the user as input and outputs the values ​​for the plurality of evaluation items as estimated values.

17. An evaluation value estimation program that causes a computer to execute an evaluation value estimation method for estimating a user evaluation value, which is a value representing the user's willingness to deal with an object, When the evaluation estimation program is loaded by the computer, an emotion recognition unit and an evaluation estimation unit are constructed within the computer. The emotion recognition unit performs the step of recognizing information on the user's emotional valence intensity and arousal level towards the object, The evaluation value estimation unit includes the step of estimating the user evaluation value using a pre-trained estimation model that takes the user's emotional valence intensity and arousal level information as input and outputs the user evaluation value, The user evaluation value includes a value indicating the degree of multiple evaluation criteria that show the type of motivation towards the object, with respect to evaluation items based on the five human senses regarding the object. The pre-trained model for estimation is a model constructed by training a predetermined model with information on emotional valence intensity and arousal level toward the object and a value indicating motivation toward the object as training data, and is characterized in that, at the time of estimation, it takes as input one piece of information on emotional valence intensity and arousal level toward the object from the user, and outputs as an estimated value a value indicating the degree of the multiple evaluation criteria relating to the evaluation item.

18. A recording medium that records the evaluation value estimation program described in claim 16 or claim 17, and is characterized in that the evaluation value estimation program is readable by the computer.

Images