Learning device, recommendation device, and methods and programs thereof

Abstract

According to the present invention, from information for specifying each musical piece of a plurality of musical pieces, biological information of a user before the user listens to the musical piece, and information relating to the improvement of motor ability of the user with listening to the musical piece, an estimation model for, with biological information as input, obtaining information relating to a musical piece estimated to have the largest effect of improving the motor ability of the user in the state of the inputted biological information or information relating to a predetermined number of musical pieces in order from the musical piece estimated to have the largest effect is learned. Using the estimation model obtained by learning, on the basis of biological information of a user that is to be recommended a musical piece, information relating to a musical piece estimated to have the largest effect of improving the motor ability when the user is in the state of the biological information or information relating to a predetermined number of musical pieces in order from the musical piece estimated to have the largest effect is obtained as a recommendation result.

Description

[Technical field]

[0001] The present invention relates to a technique for obtaining recommendation information about music pieces to be presented to a user in order to temporarily enhance the athletic ability of the user. [Background technology]

[0002] Research on the effect of listening to music on bodily activities such as the autonomic nervous system and the sympathetic nervous system of the autonomic nervous system has been conducted for some time. For example, Non-Patent Document 1 describes that listening to music changes the heart rate. Also, for example, Non-Patent Document 2 describes that listening to relaxing music causes oxytocin to be secreted in the brain, which makes people feel relaxed. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Bernardi, L., Porta, C., & Sleight, P., "Cardiovascular, cerebrovascular, and respiratory changes induced by different types of music in musicians and non-musicians: The importance of silence", Heart, 2006, 92(4), 445-452. [Non-Patent Document 2] Ooishi, Y., Mukai, H., Watanabe, K., Kawato, S., & Kashino, M., "Increase in salivary oxytocin and decrease in salivary cortisol after listening to relaxing slow-tempo and exciting fast-tempo music", PLoS ONE, 2017, 12(12), 1-16. Summary of the Invention [Problem to be solved by the invention]

[0004] Previous studies such as Non-Patent Document 1 and Non-Patent Document 2 have examined the effect of listening to music on the activity of internal organs and nerves, but no results are known about the effect of listening to music on athletic ability. The present invention aims to provide a technology that obtains recommendation information about songs to be presented to temporarily improve a user's athletic ability, taking into account the effect of listening to music on athletic ability. [Means for solving the problem]

[0005] In order to solve the above problems, according to one aspect of the present invention, the learning device includes learning data consisting of a set of learning information for each of a plurality of songs, and the set of learning information for each song is a combination of song information, which is information for identifying the song, biometric information obtained for each of the target users who are the subjects of learning before listening to the song, and an index value for each of the target users regarding improvement in athletic ability due to listening to the song, and the learning device includes a learning unit that uses the input learning data to learn an estimation model that uses the biometric information as input to obtain information on the song that is estimated to have the greatest effect on improving the athletic ability of a user in the state of the input biometric information.

[0006] In order to solve the above problem, according to another aspect of the present invention, the learning device includes a learning unit that uses the input learning data to learn an estimation model that uses the biometric information as input to obtain information on a predetermined number of songs, starting with the song that is estimated to have the greatest effect on improving the athletic ability of a user in the state of the input biometric information, and the learning data for each song is a combination of song information, which is information for identifying the song, biometric information obtained for each of the target users who are the subjects of learning before listening to the song, and an index value for each of the target users regarding improvement in athletic ability due to listening to the song, using the input learning data to learn an estimation model that uses the biometric information as input to obtain information on a predetermined number of songs, starting with the song that is estimated to have the greatest effect on improving the athletic ability of a user in the state of the input biometric information.

[0007] In order to solve the above problem, according to another aspect of the present invention, the recommendation device includes a music recommendation information generation unit that pre-stores an estimation model that uses biometric information as input to obtain music information for songs that are estimated to have the greatest effect on improving the athletic ability of a user in the state of the input biometric information, and uses the estimation model to obtain information on songs that are estimated to have the greatest effect on improving the athletic ability of the recommendation target user in the state of the input biometric information, using the biometric information of a recommendation target user, who is a user to whom songs are recommended, as input.

[0008] In order to solve the above problem, according to another aspect of the present invention, the recommendation device includes an estimation model that uses biometric information as input to obtain song information for a predetermined number of songs, starting from the song that is estimated to have the greatest effect on improving the athletic ability of a user in the state of the input biometric information, and includes a song recommendation information generation unit that uses the estimation model to input biometric information of a recommendation target user, who is a user to whom songs are recommended, to obtain information on a predetermined number of songs, starting from the song that is estimated to have the greatest effect on improving the athletic ability of the recommendation target user in the state of the biometric information. Effect of the Invention

[0009] According to the present invention, it is possible to obtain recommendation information about music to be presented to a user in order to temporarily improve the athletic ability of the user. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram showing an example of a recommendation system including a learning device and a recommendation device. [Diagram 2] FIG. 2 is a diagram illustrating an example of a processing flow of the learning device. [Diagram 3] FIG. 3 is a diagram illustrating an example of a processing flow of the recommendation device. [Figure 4] FIG. 4 is a diagram illustrating an example of the configuration of a computer that functions as at least one of a learning device and a recommendation device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] <Background of the invention> It is known that sympathetic nerves are heightened during exercise, and considering the results of the above-mentioned previous research that listening to music affects the sympathetic nerves, the hypothesis that listening to music may lead to improved athletic ability is valid. Therefore, the inventor investigated the effect of listening to music on athletic ability. Specifically, using vertical jump as a specific example of exercise, multiple subjects were asked to perform a first jump before listening to multiple songs and the height was measured, and then a second jump was performed after listening to the songs and the height was measured, and an experiment was conducted to confirm whether or not and to what extent the jump height improved. As a result of the experiment, it was confirmed that listening to music improved the jump height. In other words, it was confirmed that listening to music to subjects temporarily improved the athletic ability of the subjects. However, it was also confirmed that the improvement in jump height depended on the subjects and the songs.

[0012] Although the scientific mechanism of how music listening improves athletic ability is not yet understood, based on previous research and the experimental results confirmed this time, if a set of music information, bioinformation of the user before listening to the music, and information on the improvement of the user's athletic ability by listening to the music is acquired in advance for multiple songs, it should be possible to use the information acquired in advance to recommend what music should be presented and listened to when the user wants to temporarily improve their athletic ability based on the user's bioinformation. Therefore, the present invention learns an estimation model for recommending music from the set of music information, bioinformation of the user before listening to the music, and information on the improvement of the user's athletic ability by listening to the music acquired for multiple songs, and uses the estimation model acquired by learning to obtain recommendation information about music that is expected to temporarily improve the athletic ability of the user to whom the music is recommended, based on the bioinformation of the user to whom the music is recommended.

[0013] First Embodiment The recommendation system of the first embodiment obtains, as a recommendation result, information on music to be presented to temporarily improve the athletic ability of a user. The recommendation system 300 of the first embodiment includes a learning device 100 and a recommendation device 200, as shown in FIG.

[0014] The learning device and the recommendation device are special devices configured by loading a special program into a publicly known or dedicated computer having, for example, a central processing unit (CPU), a main storage device (RAM), and the like. The learning device and the recommendation device execute each process under the control of, for example, the central processing unit. Data input to the learning device and the recommendation device and data obtained in each process are stored in, for example, the main storage device, and the data stored in the main storage device is read out to the central processing unit as necessary and used for other processes. At least a part of each processing unit of the learning device and the recommendation device may be configured by hardware such as an integrated circuit. Each storage unit of the learning device and the recommendation device may be configured by, for example, a main storage device such as a RAM (Random Access Memory), or middleware such as a relational database or a key-value store. However, each storage unit does not necessarily have to be provided inside the learning device and the recommendation device, and may be configured by an auxiliary storage device configured by a non-transitory recording medium such as a hard disk or an optical disk, and may be configured to be provided outside the learning device and the recommendation device. The same applies to the learning device and the recommendation device of the second embodiment described later.

[0015] First, the learning device 100 will be described. [Learning device 100] Learning data is input to the learning device 100, which includes information for identifying a plurality of pieces of music, biometric information of the user before listening to the pieces of music, and information on the improvement of the user's athletic ability by listening to the pieces of music. The learning device 100 uses the input learning data to learn an estimation model that, using biometric information as input, obtains, as a recommendation result, information on a piece of music that is estimated to have the greatest effect on improving the athletic ability of the user in the state of the biometric information, and outputs the learned estimation model. As illustrated in FIG. 1, the learning device 100 includes a learning data acquisition unit 110, a learning unit 120, and a model output unit 130. The learning device 100 performs the processes of steps S110, S120, and S130 illustrated in FIG. 2.

[0016] [Learning data acquisition unit 110] The learning data acquiring unit 110 functions at least as an interface for reading into the learning device 100 learning data acquired by a device other than the learning device 100 and / or learning data stored in a device other than the learning device 100. The learning data acquired by the learning device 100 is input to the learning data acquiring unit 110. The learning data acquiring unit 110 outputs the input learning data to the learning unit 120 (step S110).

[0017] The learning data consists of a learning information set for each of a plurality of songs. The learning information set for each song is a combination of song information, bioinformation acquired for each user (hereinafter referred to as "study target user") before listening to the song, and an index value for each training target user regarding improvement in motor ability by listening to the song. For example, let j represent the song indices 1, ..., J, and k represent the user indices 1, ..., K at the time of learning. Then, the training information set S(j) for each song M(j) is a set S(j)={A(j), B(j,1), ..., B(j,K), C(j,1), ..., C(j,K)} consisting of song information A(j) for song M(j), biometric information B(j,k) for each user U(k) before listening to song M(j), and index value C(j,k) for each user U(k) regarding improvement in athletic ability due to listening to song M(j), and the training data consists of J training information sets S(1), ..., S(J).

[0018] The music information A(j) is information for identifying one of the music pieces M(j) among the multiple music pieces M(1), ..., M(J), and is, for example, the title of the music piece. The biometric information B(j,k) is biometric information of the user U(k) acquired before the user U(k) listens to the music piece M(j), and is, for example, biometric information such as an electrocardiogram, heart rate, respiration, mental sweating, and pupil diameter acquired by a sensor attached to the user U(k) a first predetermined time before the user U(k) starts listening to the music piece M(j). The index value C(j,k) relating to improvement of athletic ability is an index value relating to improvement of athletic ability due to listening to the music piece (j), obtained from a result of the user U(k) performing a predetermined exercise a second predetermined time before the user U(k) starts listening to the music piece M(j) and a result of the user U(k) performing the predetermined exercise a third predetermined time after the user U(k) finishes listening to the music piece M(j), and is, for example, a value representing the amount or degree of improvement of athletic ability due to listening to the music piece (j). For example, if the predetermined exercise is a vertical jump, and the height of the vertical jump performed by the user U(k) a second predetermined time before the user U(k) starts listening to the song M(j) is H1(k), and the height of the vertical jump performed by the user U(k) a third predetermined time after the user U(k) finishes listening to the song M(j) is H2(k), then, for example, the value H2(k)-H1(k) obtained by subtracting H1(k) from H2(k), or the value H2(k) divided by H1(k) may be used. The index value C(j,k) relating to the improvement of the athletic ability may be set to the value H2(k) / H1(k) obtained by subtracting H1(k) from H2(k), the value H2(k)-H1(k) obtained by dividing the result by H1(k), or the value H2(k)-H1(k) obtained by subtracting H1(k) from H2(k), the value H2(k)-H1(k) obtained by dividing the result by H2(k), or the value (H2(k)-H1(k)) / H2(k) obtained by subtracting H1(k) from H2(k). The first, second, and third predetermined times may be determined in advance by experiments or the like. It is preferable that the first predetermined time is longer than the second predetermined time so that the biological information B(j,k) is not affected by the exercise.

[0019] The learning data acquisition unit 110 may obtain another biometric information B'(j,k) from the input biometric information B(j,k) and output the biometric information B'(j,k) instead of the biometric information B(j,k). For example, the learning data acquisition unit 110 may obtain biometric information B'(j,k) representing the mental state of the user U(k) from the biometric information B(j,k) obtained from the user U(k) and output the biometric information B'(j,k) representing the mental state of the user U(k) instead of the biometric information B(j,k) obtained from the user U(k). The biometric information B'(j,k) is, for example, an index value of the degree of tension. For example, when the input biometric information B(j,k) is the waveform of an electrocardiogram and the amount of sweating due to mental sweating, the learning data acquisition unit 110 may obtain and output an index value of the degree of tension as the biometric information B'(j,k) from the waveform of an electrocardiogram and the amount of sweating due to mental sweating. That is, the learning data acquiring unit 110 may convert a part of the input learning data and output the converted data to the learning unit 120 as learning data.

[0020] [Learning Section 120] The learning unit 120 receives the learning data output by the learning data acquisition unit 110. The learning unit 120 uses the input learning data to learn an estimation model that receives bioinformation as input and obtains information about a piece of music that is estimated to have the greatest effect of improving the athletic ability of a user in a state of the bioinformation (step S120), and outputs the learned estimation model to the model output unit 130. A well-known learning technique may be used to learn the estimation model, and the amount of learning data may be sufficient for learning the estimation model.

[0021] The learning unit 120 may learn the estimation model using the bioinformation and index values ​​related to the improvement of athletic ability of all users at the time of learning, may learn the estimation model using the bioinformation and index values ​​related to the improvement of athletic ability of users who satisfy a predetermined condition among the users at the time of learning, or may learn the estimation model using the bioinformation and index values ​​related to the improvement of athletic ability of a specific user among the users at the time of learning. When the learning unit 120 learns using the bioinformation and index values ​​related to the improvement of athletic ability of all users, an estimation model with a low degree of dependency on the user can be obtained. When the learning unit 120 learns using the bioinformation and index values ​​related to the improvement of athletic ability of a user who satisfies a predetermined condition, an estimation model with a high degree of dependency on the user who satisfies the condition can be obtained. When the learning unit 120 learns using the bioinformation and index values ​​related to the improvement of athletic ability of a specific user, an estimation model specialized for that user can be obtained.

[0022] [Model output section 130] The model output unit 130 receives the trained estimation model output by the learning unit 120. The model output unit 130 outputs the trained estimation model that has been received as an output of the learning device 100 to the recommendation device 200 (Step S130).

[0023] Next, the recommendation device 200 will be described. [Recommendation device 200] The recommendation device 200 receives input of biometric information of a user to whom music is to be recommended (hereinafter referred to as a "recommendation target user"). Using a trained estimation model, the recommendation device 200 receives the biometric information of the recommendation target user as input, and obtains and outputs, as a recommendation result, music information of a song that is estimated to have the greatest effect of improving athletic ability when the recommendation target user is in the state of that biometric information. As illustrated in FIG. 1, the recommendation device 200 includes a biometric information acquisition unit 210, a music recommendation information generation unit 220, and a music recommendation information output unit 230. The recommendation device 200 performs the processes of steps S210, S220, and S230 illustrated in FIG. 3.

[0024] [Biometric information acquisition unit 210] The biometric information of the recommendation target user input to the recommendation device 200 is input to the biometric information acquisition unit 210. The biometric information acquisition unit 210 outputs the input biometric information to the music recommendation information generation unit 220 (Step S210).

[0025] The biometric information of the recommendation target user is, for example, an electrocardiogram, a heart rate, respiration, mental sweating, and a pupil diameter acquired by a sensor attached to the recommendation target user. In addition, when the learning data acquisition unit 110 of the corresponding learning device 100 acquires and outputs another biometric information from the input biometric information, the biometric information acquisition unit 210 also acquires and outputs another biometric information from the input biometric information, similar to the learning data acquisition unit 110. An example of another biometric information is an index value of tension. For example, when the learning data acquisition unit 110 of the corresponding learning device 100 acquires and outputs an index value of tension as biometric information from the input electrocardiogram waveform and the amount of sweating of mental sweating, the biometric information acquisition unit 210 also acquires and outputs an index value of tension as biometric information from the input electrocardiogram waveform and the amount of sweating of mental sweating. In other words, the biometric information acquisition unit 210 may convert the input biometric information and output it to the music recommendation information generation unit 220.

[0026] [Music recommendation information generating unit 220] 1, the music recommendation information generation unit 220 includes a model storage unit 225. The model storage unit 225 stores in advance a trained estimation model output by the corresponding learning device 100. The trained estimation model stored in advance in the model storage unit 225 is an estimation model that receives bioinformation as an input and obtains music information of a piece of music that is estimated to have the greatest effect on improving the athletic ability of a user in a state of the bioinformation.

[0027] The music recommendation information generating unit 220 receives the bioinformation output by the bioinformation acquiring unit 210. Using a trained estimation model pre-stored in the model storage unit 225, the music recommendation information generating unit 220 receives the bioinformation of the recommendation target user as input, obtains music information of a song estimated to have the greatest effect on improving the athletic ability of the recommendation target user in the state of the bioinformation, as music recommendation information (step S220), and outputs the obtained music recommendation information to the music recommendation information output unit 230.

[0028] The learned estimation model stored in advance in the model storage unit 225 may be obtained by any of the learning performed by the learning unit 120 of the learning device 100. That is, the learned estimation model stored in advance in the model storage unit 225 may be an estimation model learned using the biometric information of all users at the time of learning and index values ​​related to improvement of athletic ability, an estimation model learned using the biometric information of a user who satisfies a predetermined condition among the users at the time of learning and index values ​​related to improvement of athletic ability, or an estimation model learned using the biometric information of a specific user among the users at the time of learning and index values ​​related to improvement of athletic ability. However, when using an estimation model learned using the biometric information of a user who satisfies a predetermined condition among the users at the time of learning and index values ​​related to improvement of athletic ability, it is preferable to use an estimation model learned using the biometric information of a user who satisfies a condition satisfied by a recommendation target user and index values ​​related to improvement of athletic ability. Also, when using an estimation model learned using the biometric information of a specific user among the users at the time of learning and index values ​​related to improvement of athletic ability, it is preferable to use an estimation model learned using the biometric information of a user who is a recommendation target user and index values ​​related to improvement of athletic ability.

[0029] [Music recommendation information output unit 230] The music recommendation information output by the music recommendation information generating unit 220 is input to the music recommendation information output unit 230. The music recommendation information output unit 230 outputs the input music recommendation information as an output of the recommendation device 200 (Step S230).

[0030] <Second embodiment> The recommendation system of the second embodiment obtains a playlist of songs to be presented to temporarily improve the athletic ability of the user as a recommendation result. The playlist of songs is information on a plurality of songs. The recommendation system 300 of the second embodiment includes a learning device 100 and a recommendation device 200, as illustrated in FIG. 1.

[0031] First, the learning device 100 will be described. [Learning device 100] Learning data is input to the learning device 100, which includes information for identifying a plurality of songs, biometric information of the user before listening to the songs, and information on the improvement of the user's athletic ability by listening to the songs. The learning device 100 uses the input learning data to learn an estimation model that, with biometric information as input, obtains a playlist of songs that are estimated to have a large effect on improving the athletic ability of the user in the state of the biometric information as a recommendation result, and outputs the learned estimation model. As illustrated in FIG. 1, the learning device 100 includes a learning data acquisition unit 110, a learning unit 120, and a model output unit 130. The learning device 100 performs the processes of steps S110, S120, and S130 illustrated in FIG. 2.

[0032] [Learning data acquisition unit 110] The learning data acquiring unit 110 functions at least as an interface for reading into the learning device 100 learning data acquired by a device other than the learning device 100 and / or learning data stored in a device other than the learning device 100. The learning data acquired by the learning device 100 is input to the learning data acquiring unit 110. The learning data acquiring unit 110 outputs the input learning data to the learning unit 120 (step S110).

[0033] The learning data consists of a learning information set for each of a plurality of songs. The learning information set for each song is a combination of song information, bioinformation acquired for each user (hereinafter referred to as "study target user") before listening to the song, and an index value for each training target user regarding improvement in motor ability by listening to the song. For example, let j represent the song indices 1, ..., J, and k represent the user indices 1, ..., K at the time of learning. Then, the training information set S(j) for each song M(j) is a set S(j)={A(j), B(j,1), ..., B(j,K), C(j,1), ..., C(j,K)} consisting of song information A(j) for song M(j), biometric information B(j,k) for each user U(k) before listening to song M(j), and index value C(j,k) for each user U(k) regarding improvement in athletic ability due to listening to song M(j), and the training data consists of J training information sets S(1), ..., S(J).

[0034] The music information A(j) is information for identifying one of the music pieces M(j) among the multiple music pieces M(1), ..., M(J), and is, for example, the title of the music piece. The biometric information B(j,k) is biometric information of the user U(k) acquired before the user U(k) listens to the music piece M(j), and is, for example, biometric information such as an electrocardiogram, heart rate, respiration, mental sweating, and pupil diameter acquired by a sensor attached to the user U(k) a first predetermined time before the user U(k) starts listening to the music piece M(j). The index value C(j,k) relating to improvement of athletic ability is an index value relating to improvement of athletic ability due to listening to the music piece (j), obtained from a result of the user U(k) performing a predetermined exercise a second predetermined time before the user U(k) starts listening to the music piece M(j) and a result of the user U(k) performing the predetermined exercise a third predetermined time after the user U(k) finishes listening to the music piece M(j), and is, for example, a value representing the amount or degree of improvement of athletic ability due to listening to the music piece (j). For example, if the predetermined exercise is a vertical jump, and the height of the vertical jump performed by the user U(k) a second predetermined time before the user U(k) starts listening to the song M(j) is H1(k), and the height of the vertical jump performed by the user U(k) a third predetermined time after the user U(k) finishes listening to the song M(j) is H2(k), then, for example, the value H2(k)-H1(k) obtained by subtracting H1(k) from H2(k), or the value H2(k) divided by H1(k) may be used. The index value C(j,k) relating to the improvement of the athletic ability may be set to the value H2(k) / H1(k) obtained by subtracting H1(k) from H2(k), the value H2(k)-H1(k) obtained by dividing the result by H1(k), or the value H2(k)-H1(k) obtained by subtracting H1(k) from H2(k), the value H2(k)-H1(k) obtained by dividing the result by H2(k), or the value (H2(k)-H1(k)) / H2(k) obtained by subtracting H1(k) from H2(k). The first, second, and third predetermined times may be determined in advance by experiments or the like. It is preferable that the first predetermined time is longer than the second predetermined time so that the biological information B(j,k) is not affected by the exercise.

[0035] The learning data acquisition unit 110 may obtain another biometric information B'(j,k) from the input biometric information B(j,k) and output the biometric information B'(j,k) instead of the biometric information B(j,k). For example, the learning data acquisition unit 110 may obtain biometric information B'(j,k) representing the mental state of the user U(k) from the biometric information B(j,k) obtained from the user U(k) and output the biometric information B'(j,k) representing the mental state of the user U(k) instead of the biometric information B(j,k) obtained from the user U(k). The biometric information B'(j,k) is, for example, an index value of the degree of tension. For example, when the input biometric information B(j,k) is the waveform of an electrocardiogram and the amount of sweating due to mental sweating, the learning data acquisition unit 110 may obtain and output an index value of the degree of tension as the biometric information B'(j,k) from the waveform of an electrocardiogram and the amount of sweating due to mental sweating. That is, the learning data acquiring unit 110 may convert a part of the input learning data and output the converted data to the learning unit 120 as learning data.

[0036] [Learning Section 120] The learning unit 120 receives the learning data output by the learning data acquisition unit 110. The learning unit 120 uses the input learning data to learn an estimation model that takes bioinformation as input and obtains information on a plurality of songs that are estimated to have a large effect of improving the athletic ability of a user in the state of the bioinformation, and outputs the learned estimation model to the model output unit 130. Specifically, the learning unit 120 uses the input learning data to learn an estimation model that takes bioinformation as input and obtains music information on a predetermined number of songs in order from the music that is estimated to have the largest effect of improving the athletic ability of a user in the state of the bioinformation (step S120), and outputs the learned estimation model to the model output unit 130. A well-known learning technique may be used to learn the estimation model, and the amount of learning data may be sufficient for learning the estimation model.

[0037] The learning unit 120 may learn the estimation model using the bioinformation and index values ​​related to the improvement of athletic ability of all users at the time of learning, may learn the estimation model using the bioinformation and index values ​​related to the improvement of athletic ability of users who satisfy a predetermined condition among the users at the time of learning, or may learn the estimation model using the bioinformation and index values ​​related to the improvement of athletic ability of a specific user among the users at the time of learning. When the learning unit 120 learns using the bioinformation and index values ​​related to the improvement of athletic ability of all users, an estimation model with a low degree of dependency on the user can be obtained. When the learning unit 120 learns using the bioinformation and index values ​​related to the improvement of athletic ability of a user who satisfies a predetermined condition, an estimation model with a high degree of dependency on the user who satisfies the condition can be obtained. When the learning unit 120 learns using the bioinformation and index values ​​related to the improvement of athletic ability of a specific user, an estimation model specialized for that user can be obtained.

[0038] [Model output section 130] The model output unit 130 receives the trained estimation model output by the learning unit 120. The model output unit 130 outputs the trained estimation model that has been received as an output of the learning device 100 to the recommendation device 200 (Step S130).

[0039] Next, the recommendation device 200 will be described. [Recommendation device 200] The recommendation device 200 receives input of biometric information of a user to whom music is to be recommended (hereinafter referred to as a "recommendation target user"). Using a trained estimation model, the recommendation device 200 receives the biometric information of the recommendation target user as input, and obtains and outputs, as a recommendation result, a playlist of music that is estimated to have a large effect on improving athletic ability when the recommendation target user is in the state of the biometric information. As illustrated in FIG. 1, the recommendation device 200 includes a biometric information acquisition unit 210, a music recommendation information generation unit 220, and a music recommendation information output unit 230. The recommendation device 200 performs the processes of steps S210, S220, and S230 illustrated in FIG. 3.

[0040] [Biometric information acquisition unit 210] The biometric information of the recommendation target user input to the recommendation device 200 is input to the biometric information acquisition unit 210. The biometric information acquisition unit 210 outputs the input biometric information to the music recommendation information generation unit 220 (Step S210).

[0041] The biometric information of the recommendation target user is, for example, an electrocardiogram, a heart rate, respiration, mental sweating, and a pupil diameter acquired by a sensor attached to the recommendation target user. In addition, when the learning data acquisition unit 110 of the corresponding learning device 100 acquires and outputs another biometric information from the input biometric information, the biometric information acquisition unit 210 also acquires and outputs another biometric information from the input biometric information, similar to the learning data acquisition unit 110. An example of another biometric information is an index value of tension. For example, when the learning data acquisition unit 110 of the corresponding learning device 100 acquires and outputs an index value of tension as biometric information from the input electrocardiogram waveform and the amount of sweating of mental sweating, the biometric information acquisition unit 210 also acquires and outputs an index value of tension as biometric information from the input electrocardiogram waveform and the amount of sweating of mental sweating. In other words, the biometric information acquisition unit 210 may convert the input biometric information and output it to the music recommendation information generation unit 220.

[0042] [Music recommendation information generating unit 220] 1, the music recommendation information generation unit 220 includes a model storage unit 225. The model storage unit 225 stores in advance a trained estimation model output by the corresponding learning device 100. The trained estimation model stored in advance in the model storage unit 225 is an estimation model that receives bioinformation as input and obtains music information of a predetermined number of music pieces in order from the music piece that is estimated to have the greatest effect on improving the athletic ability of a user in the state of the bioinformation.

[0043] The music recommendation information generating unit 220 receives the biometric information output by the biometric information acquiring unit 210. Using a trained estimation model pre-stored in the model storage unit 225, the music recommendation information generating unit 220 receives the biometric information of the recommendation target user as input, obtains music information of a predetermined number of songs in descending order of the music estimated to have the greatest effect on improving the athletic ability of the recommendation target user in the state of the biometric information, as music recommendation information (step S220).

[0044] The learned estimation model stored in advance in the model storage unit 225 may be obtained by any of the learning performed by the learning unit 120 of the learning device 100. That is, the learned estimation model stored in advance in the model storage unit 225 may be an estimation model learned using the biometric information of all users at the time of learning and index values ​​related to improvement of athletic ability, an estimation model learned using the biometric information of a user who satisfies a predetermined condition among the users at the time of learning and index values ​​related to improvement of athletic ability, or an estimation model learned using the biometric information of a specific user among the users at the time of learning and index values ​​related to improvement of athletic ability. However, when using an estimation model learned using the biometric information of a user who satisfies a predetermined condition among the users at the time of learning and index values ​​related to improvement of athletic ability, it is preferable to use an estimation model learned using the biometric information of a user who satisfies a condition satisfied by a recommendation target user and index values ​​related to improvement of athletic ability. Also, when using an estimation model learned using the biometric information of a specific user among the users at the time of learning and index values ​​related to improvement of athletic ability, it is preferable to use an estimation model learned using the biometric information of a user who is a recommendation target user and index values ​​related to improvement of athletic ability.

[0045] [Music recommendation information output unit 230] The music recommendation information output by the music recommendation information generating unit 220 is input to the music recommendation information output unit 230. The music recommendation information output unit 230 outputs the input music recommendation information as an output of the recommendation device 200 (Step S230).

[0046] <Other examples> In the first and second embodiments, the predetermined exercise is described as a vertical jump, but the predetermined exercise is not limited to a vertical jump and may be any exercise. However, in view of the background of the invention described above, the predetermined exercise is preferably an explosive exercise such as ball throwing, throwing events, sprinting, and long jump. In other words, the recommendation system 300 of the first embodiment is expected to be particularly effective when targeting explosive exercises. That is, as an index value for improvement of athletic ability included in the learning information set handled by the learning device 100, an index value for the exercise result of the explosive exercise may be used, and more specifically, a value representing the amount or degree of improvement in the exercise result of the explosive exercise may be used. In this way, the learning device 100 can learn an estimation model that obtains information on the song that is estimated to have the greatest effect on improving the exercise results for the user's explosive exercise, or information on a predetermined number of songs starting from the song that is estimated to have the greatest effect, and the recommendation device 200 can obtain information on the song that is estimated to have the greatest effect on improving the exercise results for the explosive exercise of the user to be recommended, or information on a predetermined number of songs starting from the song that is estimated to have the greatest effect.

[0047] <Programs and recording media> The processing of each unit of the learning device and the recommendation device described above may be realized by a computer, in which case the processing contents of the functions that each device should have are described by a program. Then, by loading this program into the storage unit 1020 of the computer 1000 shown in Fig. 4 and operating the arithmetic processing unit 1010, the input unit 1030, the output unit 1040, etc., various processing functions of each of the above devices are realized on the computer.

[0048] The program describing the processing contents can be recorded on a computer-readable recording medium. The computer-readable recording medium is, for example, a non-transitory recording medium, specifically, a magnetic recording device, an optical disk, etc.

[0049] The program may be distributed, for example, by selling, transferring, lending, etc. portable recording media such as DVDs and CD-ROMs on which the program is recorded. Furthermore, the program may be distributed by storing the program in a storage device of a server computer and transferring the program from the server computer to other computers via a network.

[0050] A computer that executes such a program, for example, first stores a program recorded on a portable recording medium or a program transferred from a server computer in the auxiliary recording unit 1050, which is a non-transient storage device of the computer. Then, when executing a process, the computer reads the program stored in the auxiliary recording unit 1050, which is a non-transient storage device of the computer, into the storage unit 1020, and executes a process according to the read program. As another execution form of the program, the computer may directly read the program from the portable recording medium into the storage unit 1020 and execute a process according to the program, or may execute a process according to the received program each time a program is transferred from the server computer to the computer. Alternatively, the server computer may not transfer the program to the computer, but may execute the above-mentioned process by a so-called ASP (Application Service Provider) type service that realizes a processing function only by issuing an execution instruction and obtaining a result. Note that the program in this embodiment includes information used for processing by an electronic computer and equivalent to a program (data that is not a direct instruction to a computer but has a nature that specifies the processing of the computer, etc.).

[0051] In addition, in this embodiment, the present device is configured by executing a specific program on a computer, but at least a part of the processing contents may be realized by hardware.

[0052] Needless to say, other modifications can be made without departing from the spirit of the present invention.

Claims

1. The training data includes a set of training information for each of a plurality of pieces of music, The learning information set for each piece of music is a combination of music information, which is information for identifying the piece of music, bioinformation acquired before listening to the piece of music for each of the learning target users who are users to be studied, and an index value for each of the learning target users regarding improvement in athletic ability by listening to the piece of music; a learning unit that uses input learning data to learn an estimation model that obtains information on a piece of music that is estimated to have the greatest effect on improving the athletic ability of a user in a state of the inputted biological information, with biological information as an input; A learning device comprising:

2. The training data includes a set of training information for each of a plurality of pieces of music, The learning information set for each piece of music is a combination of music information, which is information for identifying the piece of music, bioinformation acquired before listening to the piece of music for each of the learning target users who are users to be studied, and an index value for each of the learning target users regarding improvement in athletic ability by listening to the piece of music; a learning unit that uses the input learning data to learn an estimation model that acquires information on a predetermined number of pieces of music in order from the piece of music that is estimated to have the greatest effect on improving the athletic ability of the user in the state of the inputted biological information, with biological information as an input; A learning device comprising:

3. an estimation model is stored in advance, the estimation model being configured to obtain, using biometric information as an input, music information of a piece of music that is estimated to have the greatest effect of improving the athletic ability of a user in a state where the biometric information has been input; a music recommendation information generation unit that uses the estimation model to input biometric information of a recommendation target user who is a user to whom music is to be recommended, and obtains information on music that is estimated to have the greatest effect on improving the athletic ability of the recommendation target user in the state of the biometric information. A recommendation device comprising:

4. an estimation model is stored in advance, which uses biometric information as an input and obtains music information of a predetermined number of music pieces in order from the music piece that is estimated to have the greatest effect on improving the athletic ability of a user in a state of the input biometric information; a music recommendation information generation unit that uses the estimation model to input biometric information of a recommendation target user who is a user to whom music is to be recommended, and obtains information on a predetermined number of music pieces in order from the music piece that is estimated to have the greatest effect on improving the athletic ability of the recommendation target user in the state of the biometric information; A recommendation device comprising:

5. The training data includes a set of training information for each of a plurality of pieces of music, The learning information set for each piece of music is a combination of music information, which is information for identifying the piece of music, bioinformation acquired before listening to the piece of music for each of the learning target users who are users to be studied, and an index value for each of the learning target users regarding improvement in athletic ability by listening to the piece of music; a learning step of learning an estimation model that uses the input learning data, inputs bioinformation, and obtains information on a piece of music that is estimated to have the greatest effect on improving the motor skills of a user in a state of the input bioinformation; Learning methods including:

6. The training data includes a set of training information for each of a plurality of pieces of music, The learning information set for each piece of music is a combination of music information, which is information for identifying the piece of music, bioinformation acquired before listening to the piece of music for each of the learning target users who are users to be studied, and an index value for each of the learning target users regarding improvement in athletic ability by listening to the piece of music; a learning step of learning an estimation model that uses the input learning data, inputs bioinformation, and obtains information on a predetermined number of pieces of music in order from the piece of music that is estimated to have the greatest effect on improving the athletic ability of the user in the state of the input bioinformation; Learning methods including:

7. an estimation model is stored in advance, the estimation model being configured to obtain, using biometric information as an input, music information of a piece of music that is estimated to have the greatest effect of improving the athletic ability of a user in a state of the input biometric information; a music recommendation information generating step of using the estimation model to input biometric information of a recommendation target user who is a user to whom music is to be recommended, and obtaining information on music that is estimated to have the greatest effect on improving the athletic ability of the recommendation target user in the state of the biometric information; Including,recommended methods.

8. an estimation model is stored in advance, which uses biometric information as an input and obtains music information of a predetermined number of music pieces in order from the music piece that is estimated to have the greatest effect on improving the athletic ability of a user in a state of the input biometric information; a music recommendation information generating step of using the estimation model to input biometric information of a recommendation target user who is a user to whom music is to be recommended, and obtaining information on a predetermined number of music pieces in order from the music piece estimated to have the greatest effect on improving the athletic ability of the recommendation target user in the state of the biometric information; Including,recommended methods.

9. A program for causing a computer to function as the learning device according to claim 1 or 2.

10. A program for causing a computer to function as the recommendation device according to claim 3 or 4.

Images