JP6946929B2 - Recommendation systems, devices, methods, and programs - Google Patents

Description

The present invention relates to recommendation systems, devices, methods, and programs.

In recent years, an online market that provides various applications (hereinafter referred to as "applications") to smart devices has been used. Users of smart devices can download desired apps from the market for a fee or for free and use them on their smart devices.

Conventionally, as a method of improving the operability of an application, a method of recommending an application setting value useful for the user has been proposed. As a method of recommending the set value, a method of extracting the difference between the set value of the target job and the executed job similar to the target job and recommending the set value extracted as the difference has been proposed.

However, in the above-mentioned conventional method, since the setting value different between similar jobs is only recommended, it is unclear whether the recommended setting value is a useful setting value for the user.

The present invention has been made in view of the above problems, and an object of the present invention is to make it possible to recommend setting values useful to the user.

The recommendation system according to one embodiment includes a classification unit that classifies the usage tendency of applications of a plurality of users into a plurality of clusters, the usefulness of the setting value of the application in the cluster to which the usage tendency of the target user belongs, and a recommendation. A recommendation unit that recommends the set value of the application to the target user based on the conditions, and a learning unit that updates at least one of the classification method by the classification unit and the recommendation condition based on the usage tendency. , Equipped with.

According to each embodiment of the present invention, setting values useful for the user can be recommended.

The figure which shows an example of the structure of a recommendation system. The figure which shows an example of the hardware configuration of a user terminal. The figure which shows an example of the functional structure of a recommendation system. The figure which shows an example of the operation log. The figure which shows an example of the usage tendency stored in the usage tendency storage part. The figure which shows an example of the usage tendency. The figure which shows an example of the usage tendency. The figure which shows an example of the usefulness. The figure which shows the success example and the failure example of the recommendation. A sequence diagram showing an example of usage tendency and cluster update processing. A sequence diagram showing an example of recommendation processing of set values. A sequence diagram showing an example of usefulness update processing. The sequence diagram which shows an example of the learning process of a classification method and a recommendation condition.

Hereinafter, each embodiment of the present invention will be described with reference to the accompanying drawings. Regarding the description of the specification and the drawings according to each embodiment, the components having substantially the same functional configuration are designated by the same reference numerals, and the superimposed description will be omitted.

The recommendation system 1000 according to one embodiment will be described with reference to FIGS. 1 to 13. First, the configuration of the recommendation system 1000 will be described. FIG. 1 is a diagram showing an example of the configuration of the recommendation system 1000. The recommendation system 1000 of FIG. 1 includes user terminals 1A to 1B, a recommendation server 2, an application market 3, and a network 4.

The user terminals 1A to 1C are computers for the user to use various applications. User terminals 1A to 1C are, for example, PCs (Personal Computers), smartphones, tablet terminals, smart watches, or televisions, but are not limited thereto. Hereinafter, when the user terminals 1A to 1C are not distinguished, they are simply referred to as the user terminal 1. In the example of FIG. 1, the recommendation system 1000 includes three user terminals 1, but the number of user terminals 1 included in the recommendation system 1000 is arbitrary. Each application that can be used on the user terminal 1 has a plurality of application settings, and each application setting has a plurality of setting values that can be set as default values.

The recommendation server 2 is an example of a recommendation device, and is a server computer that recommends to the user of the user terminal 1 the setting value of the application, which is useful as a default value for the user. The recommendation server 2 recommends a setting value useful as a default value to the user of each user terminal 1 based on the information acquired from each user terminal 1 included in the recommendation system 1000. In this embodiment, the recommendation device may be, for example, a PC, a smartphone, a tablet terminal, or the like. Further, the recommendation device may be composed of a plurality of server computers.

The application market 3 is an online market that provides users with applications for a fee or for free, and is composed of one or a plurality of server computers. The user can download the desired application from the application market 3 to the user terminal 1.

The network 4 includes a LAN (Local Area Network), the Internet, and the like, and connects the user terminal 1, the recommendation server 2, and the application market 3 to each other.

Next, the hardware configuration of the user terminal 1 according to the present embodiment will be described. FIG. 2 is a diagram showing an example of the hardware configuration of the user terminal 1. The user terminal 1 of FIG. 2 includes a CPU (Central Processing Unit) 101, a ROM (Read Only Memory) 102, a RAM (Random Access Memory) 103, and an HDD (Hard Disk Drive) 104. Further, the user terminal 1 includes an input device 105, a display device 106, a communication interface 107, and a system bus 108.

By executing the program, the CPU 101 controls each configuration of the user terminal 1 and realizes the function of the user terminal 1.

The ROM 102 stores a program (such as a BIOS (Basic Input Output System)) executed by the CPU 101 and various data.

The RAM 103 provides the CPU 101 with a work area.

The HDD 104 stores a program (OS (Operating System) or the like) executed by the CPU 101 and various data.

The input device 105 inputs information according to the user's operation to the user terminal 1. The input device 105 is, for example, a touch panel, a keyboard, a mouse, or a hardware button, but is not limited thereto.

The display device 106 displays a screen according to the operation of the user. The display device 106 is, for example, a liquid crystal display, an organic EL (Electro Luminescence) display, or a plasma display, but is not limited thereto.

The communication interface 107 is an interface for connecting the user terminal 1 to the network 4. The user terminal 1 communicates with the recommendation server 2 and the application market 3 connected to the network 4 via the communication interface 107.

The system bus 108 connects the CPU 101, the ROM 102, the RAM 103, the HDD 104, the input device 105, the display device 106, and the communication interface 107 to each other.

Similar to the user terminal 1, the recommendation server 2 includes a CPU 201, a ROM 202, a RAM 203, an HDD 204, an input device 205, a display device 206, and a communication interface 207, which provide a system bus 208. They are connected to each other via. Since each hardware is the same as that of the user terminal 1, the description thereof will be omitted.

Next, the functional configuration of the recommendation system 1000 according to the present embodiment will be described. FIG. 3 is a diagram showing an example of the functional configuration of the recommendation system 1000.

First, the functional configuration of the user terminal 1 will be described. The user terminal 1 of FIG. 3 includes a terminal control unit 11, an application storage unit 12, an operation log storage unit 13, and an upload unit 14. Each of these functional configurations is realized by the CPU 101 executing a program and cooperating with each hardware of the user terminal 1.

The terminal control unit 11 controls the overall operation of the user terminal 1. Specifically, the terminal control unit 11 downloads the application from the application market 3, installs and uninstalls the application, updates the operation log, recommends the application to the user, and the like. The operation of the terminal control unit 11 will be described in detail later.

The application storage unit 12 stores the application (program) installed in the terminal control unit 11 and the application information of the application. The application information includes information such as an application name and an application ID. The application ID is the identification information of the application.

The operation log storage unit 13 stores the operation log of the user terminal 1 by the user. The operation log is data in which the execution date and time of the operation and the operation content are associated with each other. The operation contents include download, installation, uninstallation, start, end, screen switching, setting value change, job execution, etc. of the application.

FIG. 4 is a diagram showing an example of an operation log stored in the operation log storage unit 13. Each record in FIG. 4 corresponds to an operation log corresponding to each operation. In the example of FIG. 4, the download of the application A is executed at 15:00 on February 1, and the installation of the application A is executed at 15:01 on February 1. Further, after the application A is started, the job is executed after switching the screen and changing the default value a1 of the setting a to the setting value a2 (changing the setting value).

When the user terminal 1 is used by a plurality of users, the operation log corresponding to each operation may include the user ID of the user who executed the operation. The user ID is user identification information. Further, instead of including the user ID in each operation log, the operation content in which the operation log is stored may include the login of the user. In this case, the operation log corresponding to the user's login includes the execution date and time (login date and time) and the user ID of the logged-in user. Hereinafter, the user whose user ID is X will be referred to as user X.

The upload unit 14 recommends the operation log stored in the operation log storage unit 13 periodically or at a predetermined timing (such as when the operation log is updated or when the power is turned on to the user terminal 1). Upload to.

Next, the functional configuration of the recommendation server 2 will be described. The recommendation server 2 of FIG. 3 includes a recommendation control unit 21, a usage tendency storage unit 22, a setting information storage unit 23, a classification unit 24, a recommendation unit 25, and a learning unit 26. Each of these functional configurations is realized by the CPU 201 executing a program and cooperating with each hardware of the recommendation server 2.

The recommendation control unit 21 controls the overall operation of the recommendation server 2. Specifically, the recommendation control unit 21 receives the operation log uploaded from the user terminal 1, updates the usage tendency, and requests the operation of each function configuration. In addition, the recommendation control unit 21 acquires setting information from the user terminal 1. The operation of the recommendation control unit 21 will be described in detail later.

The usage tendency storage unit 22 stores the usage tendency of each application of each user for each predetermined period. The predetermined period can be arbitrarily set, such as one day, one week, and one month. In addition, the usage tendency includes the number of uses (the number of startups or the number of job executions) and the average number of steps.

The average number of steps is the average value of the number of steps (number of operations) from the start of the application to the execution of the job. For example, in the example of FIG. 4, the job of the application A is executed through three operations of switching the screen, changing the set value, and executing the job after the application A is started, so that the number of steps is three. The number of steps corresponds to the amount of effort required for the user to execute the job, and the smaller the number of steps, the less the effort of the user. From this, it is considered that the default value of the application having a small average number of steps is a useful default value for the user, which reduces the time and effort of the user.

Further, the usage tendency storage unit 22 stores the cluster to which the usage tendency of each user belongs for each predetermined period. As will be described later, the usage tendency of each user for each predetermined period is classified into a plurality of clusters by the classification unit 24. The usage tendency storage unit 22 stores the clusters classified by the classification unit 24 in association with the usage tendency of each user for each predetermined period.

FIG. 5 is a diagram showing an example of the usage tendency stored in the usage tendency storage unit 22. In the example of FIG. 5, the weekly usage tendency (number of uses and number of steps) of the applications A and B by the users U001 and U002 is stored in association with the cluster. For example, the number of times application A is used by user U001 from February 1st to February 7th is 50 times, the average number of steps is 5.0, the number of times application B is used 100 times, and the average number of steps is 4.0. Yes, the user U001's application usage tendency belongs to cluster 1. In FIG. 5, the value “−” of the number of times the application is used indicates that the application is not installed. As an application that remembers the number of times of use even though it is not installed, an application that is recommended but not installed can be considered.

The usage tendency stored in the usage tendency storage unit 22 is updated by the recommendation control unit 21 based on, for example, the operation log uploaded from each user terminal 1. Specifically, the recommendation control unit 21 aggregates the usage tendency of each application for each predetermined period based on the operation log uploaded from the user terminal 1, and calculates the usage tendency of each application of the user of the user terminal 1. Add to usage trends. For example, when the recommendation control unit 21 stores the usage tendency of FIG. 5 in the usage tendency storage unit 22 and uploads the operation log of FIG. 4 from the user terminal 1, the recommendation control unit 21 starts from February 1 to 2 of the user of the user terminal 1. Add 1 each for the number of times apps A and B are used until the 6th of the month. In addition, the recommendation control unit 21 updates the average number of steps of the application A during the period from February 1st to February 6th of the user of the user terminal 1.

Further, the upload unit 14 of the user terminal 1 aggregates the usage tendency of each application for each predetermined period based on the operation log stored in the operation log storage unit 13, and uploads the aggregated usage tendency to the recommendation server 2. You may. In this case, the recommendation control unit 21 may add the uploaded usage tendency to the usage tendency stored in the usage tendency storage unit 22 as it is.

The setting information storage unit 23 stores the setting information of each user terminal 1. The setting information includes the terminal ID and address of the user terminal 1, the application ID of the installed application, the default value being set, and the like. The terminal ID is the identification information of the user terminal 1. The default value is stored as history information in association with the set date and time. The recommendation control unit 21 acquires setting information from the user terminal 1 periodically or at a predetermined timing (such as when the user terminal 1 connects to the recommendation server 2 or when an operation log is acquired from the user terminal 1). The setting information stored in the setting information storage unit 23 is updated.

The classification unit 24 periodically or at a predetermined timing (such as when the usage tendency is updated), is based on the usage tendency of a plurality of users stored in the usage tendency storage unit 22, and is used for each user at a predetermined period. Classify usage trends into multiple clusters. A cluster is a collection of similar usage trends. The clusters classified by the classification unit 24 are stored in the usage tendency storage unit 22 in association with the usage tendency belonging to the cluster.

The classification unit 24 stores a classification method for classifying usage trends. The classification unit 24 can use any clustering method such as the shortest distance method, Ward's method, and k-means method as the method of classifying the usage tendency. In this case, the classification unit 24 may calculate the distance (dissimilarity) between usage tendencies based on the value of the usage tendency for each application, or for each application based on the value of the usage tendency for each application. The usage tendency ratio may be calculated and calculated based on the usage tendency ratio for each application.

6 and 7 are diagrams showing an example of usage tendency. In the example of FIG. 6, the number of times of use of the applications A to D of the users U001 to U003 is shown. In the example of FIG. 6, the values of the number of times of use of the user U001 and the user U002 are similar, and the ratio of the number of times of use is similar between the user U002 and the user U003.

Further, in the example of FIG. 7, the average number of steps of the applications A to D of the users U001 to U003 is shown. In the example of FIG. 7, the values of the average number of steps of the user U002 and the user U003 are similar, but the values of the average number of steps of the user U001 are not similar.

In the example of FIG. 6, when the classification unit 24 calculates the distance between usage tendencies based on the usage tendency values, the usage tendencies of users U001 and U002 are classified into the same cluster, and the usage tendencies of users U003 are different clusters. are categorized. On the other hand, when the classification unit 24 calculates the distance between usage tendencies based on the usage tendency ratio, the usage tendencies of users U002 and U003 are classified into the same cluster, and the usage tendencies of user U001 are classified into different clusters. ..

Further, the classification unit 24 may use a method according to a predetermined classification condition as a method for classifying the usage tendency. For example, the classification unit 24 may classify the usage tendency for each application having the largest usage tendency value or ratio, or classify the usage tendency for each application whose usage tendency value or ratio is equal to or higher than the threshold value. May be good. In the case of the former classification condition, in the example of FIG. 6, the usage tendencies of the users U002 and U003 are classified into the same cluster, and in the example of FIG. 7, the usage tendencies of the users U001 to U003 are classified into the same cluster.

The recommendation unit 25 selects a setting value useful as a default value for the user at regular intervals or at a predetermined timing (such as a usage tendency or a timing when the cluster is updated), and recommends the selected setting value to the user. Specifically, the recommendation unit 25 transmits information indicating a setting value useful as a default value for the user to the user terminal 1 of the user.

The recommendation unit 25 stores the usefulness of each set value for each cluster of usage tendency. The usefulness of a certain setting value in a certain cluster is an index showing the degree of usefulness when a user belonging to the cluster has a usage tendency of setting the setting value as a default value. The method of calculating the usefulness will be described later.

FIG. 8 is a diagram showing an example of the usefulness stored in the recommendation unit 25. In the example of FIG. 8, the usefulness of each of the setting values of the settings a to c of the application A in the clusters 1 to 3 is shown. For example, the usefulness of the set value a1 in the cluster 1 is 9, the usefulness of the set value a2 is 4, and the usefulness of the set value a3 is 2. This means that the setting value a1 is the most useful as the default value of the setting a of the application A for the user whose usage tendency belongs to the cluster 1.

In addition, the recommendation unit 25 stores the recommendation conditions. The recommendation condition corresponds to the selection condition of the setting value (useful setting value) recommended to the user. The recommendation conditions include that the usefulness is maximum and that the usefulness is equal to or higher than the threshold value.

The recommendation unit 25 recommends a setting value in which the usefulness satisfies the recommendation condition in the cluster as a useful setting value (useful default value) as a default value for a user belonging to the cluster having a tendency to use the cluster. For example, when the recommendation condition has a usefulness of 7 or more, in the example of FIG. 8, the recommendation unit 25 recommends the set values a1 and c1 as useful default values to the users whose usage tendency belongs to the cluster 1.

It is preferable that the recommendation unit 25 does not recommend the set value even if the usefulness is a set value satisfying the recommendation condition, if the set value has already been set as the default value. As a result, it is possible to suppress unnecessary recommendations to the user. The recommendation unit 25 can determine whether or not the set value has been set as the default value by referring to the default value of the user terminal 1 stored in the setting information storage unit 23.

The learning unit 26 determines the method of classifying the usage tendency and the set value based on the usage tendency stored in the usage tendency storage unit 22 periodically or at a predetermined timing (such as the usage tendency or the timing when the cluster is updated). Update the recommendation conditions to increase the probability of success of the recommendation. That is, the learning unit 26 learns the classification method and the recommendation condition having a high probability of success of the recommendation.

Updating the classification method means updating the parameters of the clustering method (k in the k-means method, etc.) and the parameters of the classification conditions (usage tendency value, ratio threshold value, etc.). Updating the recommendation condition means updating the parameters of the recommendation condition (such as the threshold of usefulness). The success probability of the recommendation is the probability that the recommendation of the set value by the recommendation unit 25 succeeds. The success of the recommendation means that the set value recommended by the recommendation unit 25 is set as the default value by the user, and the usage tendency of the application corresponding to the recommended set value satisfies a predetermined condition. As predetermined conditions, for example, it is conceivable that the number of times the application is used increases, the average number of steps decreases, and the like.

FIG. 9 is a diagram illustrating examples of success and failure of recommendations. 9 (A) shows an example of the number of times of use before the recommendation of user X, and FIGS. 9 (B) to 9 (D) show an example of the number of times of use after the recommendation of user X. 9 (B) and 9 (C) correspond to a failure example of the recommendation, and FIG. 9 (D) correspond to a successful example of the recommendation. In the example of FIG. 9, the default value of the setting a of the application A before the recommendation of the user X is a1, and the recommendation server 2 recommends the setting value a2 to the user X as the default value of the setting a. .. Further, the above-mentioned predetermined condition is assumed to be an increase in the number of times the application is used.

In the example of FIG. 9B, the user X does not change the default value of the setting a in response to the recommendation from the recommendation server 2. In this case, it is considered that the user has determined that the set value a2 is not useful as the default value, so the learning unit 26 determines that the recommendation to the user X has failed.

In the example of FIG. 9C, the user X changes the default value of the setting a to the setting value a2 according to the recommendation from the recommendation server 2, but the number of times the application A is used after the recommendation is reduced. .. In this case, although the user determines that the set value a2 is useful as the default value, it is considered that the set value a2 is not actually useful. Therefore, the learning unit 26 determines that the recommendation to the user X is a failure.

In the example of FIG. 9D, the user X changes the default value of the setting a to the setting value a2 according to the recommendation from the recommendation server 2, and the number of times the application A is used after the recommendation increases. There is. In this case, the user determines that the set value a2 is useful as the default value, and it is considered that the set value a2 is actually useful. Therefore, the learning unit 26 determines that the recommendation to the user X is successful.

In this way, the learning unit 26 can determine the success or failure of the recommendation based on the usage tendency before and after the recommendation and the default value before and after the recommendation. The learning unit 26 can acquire the usage tendency before and after the recommendation by referring to the usage tendency stored in the usage tendency storage unit 22. Further, the learning unit 26 can acquire the default value before and after the recommendation by referring to the setting information stored in the setting information storage unit 23.

In the present embodiment, the recommendation unit 25 recommends a set value to the user, and the user sets (or does not set) the recommended set value as a default value. Further, the upload unit 14 transmits the operation log of the application to the recommendation server 2, and the usage tendency storage unit 22 stores the usage tendency of the application aggregated based on the operation log received from the user terminal 1. Further, the learning unit 26 uses the classification method and the recommendation unit 25 used by the classification unit 24 based on the usage tendency of the application stored in the usage tendency storage unit 22 and the default value stored in the setting information storage unit 23. Learn the recommendation conditions used by. In this way, the learning unit 26 feeds back the recommendation result (default value and usage tendency after the recommendation) of the application, and learns the classification method and the recommendation condition based on the recommendation result. Therefore, the learning unit 26 can learn the classification method and the recommendation condition so as to increase the success probability of the recommendation.

The learning unit 26 can learn a classification method and a recommendation condition having a high probability of success of recommendation by using an arbitrary learning method such as a support vector machine (SVM), clustering, a neural network, and a decision tree. The learning unit 26 preferably includes a machine learning engine for learning the classification method and the recommendation condition.

Further, the learning unit 26 stores the usage tendency stored in the usage tendency storage unit 22 and the setting information storage unit 23 at regular intervals or at a predetermined timing (usage tendency, timing when the cluster is updated, etc.). Update the usefulness of each setting value for each cluster based on the default value. Here, the method of calculating the usefulness will be described by taking the usefulness of the set value a1 in the cluster X as an example. It is assumed that the setting value a1 is the setting value of the setting a of the application A.

The learning unit 26 calculates the usefulness so that the larger the average value of the number of times (or the ratio thereof) the users belonging to the cluster X use the application A, the higher the usefulness of the set value a1 in the cluster X. do. Further, the learning unit 26 is useful so that the smaller the average value of the average number of steps (or the ratio thereof) of the application A by the users whose usage tendency belongs to the cluster X, the higher the usefulness of the set value a1 in the cluster X. Calculate the degree. This is because the larger the average value of the number of times the application A is used, or the smaller the average value of the average number of steps of the application A, the more the users belonging to the cluster X are accustomed to using the application A, which is useful based on the usage tendency of the user. This is because it is considered that the reliability of the degree will be high.

Further, the learning unit 26 determines the usefulness so that the larger the ratio of the set value a1 in the default value of the application A of the user whose usage tendency belongs to the cluster X, the higher the usefulness of the set value a1 in the cluster X. calculate. This is because it is considered that the more users use the set value a1 as the default value, the higher the possibility that the set value a1 is a useful default value.

Further, in the learning unit 26, when the set value a1 is recommended to the user whose usage tendency belongs to the cluster X and then the user sets the set value a1 as the default value, the usefulness of the set value a1 in the cluster X is high. The usefulness may be calculated so as to be. This is because when the recommended setting value a1 is set to the default value, it is considered that the setting value a1 is determined by the user to be a useful default value.

Further, the learning unit 26 has a usefulness so that when a user whose usage tendency belongs to the cluster X changes the default value from the set value a1 to another set value, the usefulness of the set value a1 in the cluster X becomes low. May be calculated. This is because it is considered that when the default value is changed from the set value a1 to another set value by the user, the user determines that the set value a1 is not useful as the default value.

Further, the learning unit 26 sets the setting in the cluster X when the number of times (or the ratio thereof) of the user's application A is used increases after the default value is set to the set value a1 by the user whose usage tendency belongs to the cluster X. The usefulness may be calculated so that the usefulness of the value a1 becomes high. This is because it is considered that the number of times (or the ratio thereof) of using the application A has increased as a result of improving the convenience of the application A by setting the set value a1 to the default value.

Further, when the average number of steps (or the ratio thereof) of the application A of the user decreases after the default value is set to the set value a1 by the user whose usage tendency belongs to the cluster X, the learning unit 26 in the cluster X The usefulness may be calculated so that the usefulness of the set value a1 becomes high. This is because it is considered that by setting the setting value a1 to the default value, the time and effort required for the user to execute the job of the application A is reduced.

Next, the processing of the recommendation system 1000 according to the present embodiment will be described. FIG. 10 is a sequence diagram showing an example of usage tendency and cluster update processing.

When the update timing arrives, the upload unit 14 of the user terminal 1 requests the operation log storage unit 13 for the operation log to be transmitted at the timing (step S101). Specifically, the upload unit 14 requests the operation log storage unit 13 for the operation log (the operation log whose execution date and time is after the previous transmission timing) stored in the operation log storage unit 13 after the previous update timing. ..

When the operation log is requested, the operation log storage unit 13 passes the requested operation log to the upload unit 14 (step S102).

When the upload unit 14 receives the operation log, the upload unit 14 uploads the received operation log to the recommendation server 2 (step S103).

When the upload unit 14 uploads the operation log, the recommendation control unit 21 receives the uploaded operation log and aggregates the usage tendency of each application of each user for each predetermined period based on the operation log (step S104). ..

When the recommendation control unit 21 aggregates the usage trends, the recommendation control unit 21 stores the aggregated usage trends in the usage trend storage unit 22 (step S105). As a result, the usage tendency stored in the usage tendency storage unit 22 is updated.

Further, the recommendation control unit 21 requests the classification unit 24 to classify the usage tendency (step S106). In the example of FIG. 10, it is assumed that the classification unit 24 classifies the usage tendency according to the request from the recommendation control unit 21, but the classification unit 24 has a timing independent of the recommendation control unit 21. You may classify the usage tendency with.

When the classification unit 24 is requested to classify the usage tendency, the classification unit 24 requests the usage tendency storage unit 22 for the usage tendency of each application of each user (step S107). At this time, the classification unit 24 may request only the usage tendency (the usage tendency updated in step S107) during the period including the update timing. As a result, the load on the recommendation server 2 can be reduced and the processing speed can be increased.

Further, the classification unit 24 may request the usage tendency of a plurality of periods including the period including the update timing. Thereby, the usage tendency of a plurality of periods can be classified by the same classification method.

When the usage tendency is requested, the usage tendency storage unit 22 passes the requested usage tendency to the classification unit 24 (step S108).

Upon receiving the usage tendency, the classification unit 24 classifies the usage tendency of each user for each predetermined period based on the received usage tendency (step S109). The method of classifying the usage tendency is as described above.

When the usage tendency is classified, the classification unit 24 stores the clusters obtained by the classification in the usage tendency storage unit 22 (step S110). As a result, the cluster stored in the usage tendency storage unit 22 is updated.

By the above processing, the recommendation server 2 can update the usage tendency and the cluster based on the latest operation log. As a result, the usage tendency and cluster as shown in FIG. 5 are stored in the usage tendency storage unit 22.

FIG. 11 is a sequence diagram showing an example of the recommendation processing of the set value. Hereinafter, the user who is the target of the recommendation of the set value will be referred to as the target user. The target user may be one person or a plurality of target users.

When the recommendation timing arrives, the recommendation control unit 21 of the recommendation server 2 requests the recommendation unit 25 to recommend the set value to the target user (step S201). In the example of FIG. 11, it is assumed that the recommendation unit 25 recommends a set value in response to a request from the recommendation control unit 21, but the recommendation unit 25 is set at a timing independent of the recommendation control unit 21. You may recommend the value.

When the recommendation unit 25 is requested to recommend the set value, the recommendation unit 25 requests the usage tendency storage unit 22 for a cluster to which the usage tendency of the target user belongs during the period including the recommendation timing (step S202).

When the usage tendency storage unit 22 requests a cluster, the usage tendency storage unit 22 passes the requested cluster to the recommendation unit 25 (step S203).

When the recommendation unit 25 receives the cluster, it refers to the usefulness and recommendation condition of each setting value in the received cluster (cluster of the target user), and sets the setting value whose usefulness satisfies the recommendation condition as a useful setting as a default value. Select as a value (step S204).

When the setting value is selected, the recommendation unit 25 transmits information indicating the selected setting value to the user terminal 1 of the target user (step S206). The information indicating the setting value includes the setting value and the information indicating the application and the setting corresponding to the setting value.

When the recommendation unit 25 transmits the information indicating the set value, the terminal control unit 11 of the user terminal 1 of the target user receives the information and saves the operation log in the operation log storage unit 13 (step S206). The operation content of the operation log saved here is a recommendation of the set value.

Further, the terminal control unit 11 recommends the setting value selected by the recommendation unit 25 as a setting value useful as a default value for the target user based on the received information (step S207). The method of recommending the set value by the terminal control unit 11 is arbitrary. The terminal control unit 11 may recommend the set value by displaying a pop-up or a dialog on the display device 106 of the user terminal 1, or may recommend the set value by a push notification or an e-mail transmission.

By the above processing, the recommendation server 2 can recommend a setting value useful as a default value for the target user. Further, the usage tendency and the cluster as shown in FIG. 5 are stored in the usage tendency storage unit 22.

FIG. 12 is a sequence diagram showing an example of the usefulness update process.

When the recommendation control unit 21 of the recommendation server 2 arrives at the update timing of the usefulness, the recommendation control unit 21 requests the learning unit 26 to update the usefulness (step S401). In the example of FIG. 12, it is assumed that the learning unit 26 updates the usefulness in response to the request from the recommendation control unit 21, but the learning unit 26 is useful at a timing independent of the recommendation control unit 21. You may update the degree.

When the learning unit 26 is requested to update the usefulness, the learning unit 26 requests the usage tendency storage unit 22 for the usage tendency and the cluster (step S402). The learning unit 26 may request a usage tendency and a cluster according to the method of calculating the usefulness. The learning unit 26 requests, for example, usage trends and clusters for a plurality of periods including update timing.

When the usage tendency and cluster are requested, the usage tendency storage unit 22 passes the requested usage tendency and cluster to the learning unit 26 (step S403).

Further, the learning unit 26 requests the setting information storage unit 23 for the default value being set (step S404). The learning unit 26 may request a default value according to the method of calculating the usefulness. The learning unit 26 requests, for example, default values for a plurality of periods including update timing.

When the setting information storage unit 23 is requested for the default value, the setting information storage unit 23 passes the requested default value to the learning unit 26 (step S405).

When the learning unit 26 receives the usage tendency, the cluster, and the default value, the learning unit 26 calculates the usefulness of each set value for each cluster based on the received usage tendency, the cluster, and the default value (step S406). The method of calculating the usefulness is as described above.

When the learning unit 26 calculates the usefulness, the calculated usefulness is passed to the recommendation unit 25 (step S407). When the recommendation unit 25 receives the usefulness, it stores the received usefulness and deletes the original usefulness. As a result, the usefulness stored in the recommendation unit 25 is updated. After that, the recommendation unit 25 selects a useful setting value by utilizing the updated usefulness.

By the above processing, the usefulness can be updated. By updating the usefulness level as needed, the recommendation unit 25 can recommend to the user useful setting values according to the latest usage tendency and default value.

FIG. 13 is a sequence diagram showing an example of learning processing of the classification method and the recommendation condition.

When the learning timing of the classification method and the recommendation condition arrives, the recommendation control unit 21 of the recommendation server 2 requests the learning unit 26 to learn the classification method and the recommendation condition (step S501). In the example of FIG. 13, it is assumed that the learning unit 26 learns the classification method and the recommendation condition in response to the request from the recommendation control unit 21, but the learning unit 26 is independent of the recommendation control unit 21. You may learn the classification method and the recommendation condition at the timing.

When the learning unit 26 is requested to learn the classification method and the recommendation condition, the learning unit 26 requests the usage tendency storage unit 22 for the usage tendency and the cluster (step S502). The learning unit 26 may request a usage tendency and a cluster according to the classification method and the learning method of the recommendation condition.

When the usage tendency and cluster are requested, the usage tendency storage unit 22 passes the requested usage tendency and cluster to the learning unit 26 (step S503).

Further, the learning unit 26 requests a default value from the setting information storage unit 23 (step S504). The learning unit 26 may request a default value according to the classification method and the learning method of the recommendation condition.

When the setting information storage unit 23 is requested for the default value, the setting information storage unit 23 passes the requested default value to the learning unit 26 (step S505).

When the learning unit 26 receives the usage tendency, the cluster, and the default value, the learning unit 26 learns the classification method and the recommendation condition (parameter) having a high probability of success of the recommendation based on the received usage tendency, the cluster, and the default value. Step S506). The classification method and the learning method of the recommendation condition are as described above.

When the learning unit 26 learns the classification method, the learning unit 26 passes the learned classification method to the classification unit 24 (step S507). When the classification unit 24 receives the classification method, the classification unit 24 stores the received classification method and deletes the original classification method. As a result, the classification method stored in the classification unit 24 is updated. After that, the classification unit 24 classifies the usage tendency by using the updated classification method.

Further, when the learning unit 26 learns the recommendation condition, the learning unit 26 passes the learned recommendation condition to the recommendation unit 25 (step S508). When the recommendation unit 25 receives the recommendation condition, the recommendation unit 25 stores the received recommendation condition and deletes the original recommendation condition. As a result, the recommendation condition stored in the recommendation unit 25 is updated. After that, the recommendation unit 25 recommends a useful set value by using the updated recommendation condition.

By the above processing, the classification method and the recommendation condition can be updated. By updating the classification method as needed, the classification unit 24 can classify the user's usage tendency by the latest usage tendency and the classification method according to the default value. Further, by updating the recommendation condition at any time, the recommendation unit 25 can recommend a useful set value according to the latest usage tendency and the recommendation condition according to the default value.

As described above, the recommendation system 1000 according to the present embodiment sets useful setting values for a user who belongs to the same cluster as the target user, that is, a user who has a similar usage tendency to the target user. I recommend it to you. Since there is a high possibility that useful setting values are the same among users with similar usage tendencies, the recommendation system 1000 can recommend setting values that are useful for the target user.

In the present embodiment, at least a part of the user terminal 1 included in the recommendation system 1000 has the function of the recommendation server 2, and the user terminal 1 having the function of the recommendation server 2 serves as the recommendation device and is another user terminal 1. You may recommend a setting value that is useful for.

Further, each function of the embodiment described above can be realized by one or a plurality of processing circuits. Here, the "processing circuit" in the present specification is a processor programmed to execute each function by software such as a processor implemented by an electronic circuit, or a processor designed to execute each function described above. It shall include devices such as ASIC (Application Specific Integrated Circuit), DSP (Digital Signal Processor), FPGA (Field Programmable Gate Array) and conventional circuit modules.

Further, the present invention is not limited to the configurations shown here, such as combinations with other elements in the configurations and the like described in the above embodiments. These points can be changed without departing from the spirit of the present invention, and can be appropriately determined according to the application form thereof.

1: User terminal 2: Recommendation server 3: App market 11: Terminal control unit 12: App storage unit 13: Operation log storage unit 14: Upload unit 21: Recommendation control unit 22: Usage tendency storage unit 23: Setting information storage unit 24 : Classification unit 25: Recommendation unit 26: Learning unit 1000: Recommendation system

Japanese Unexamined Patent Publication No. 2016-007724

Claims (15)

A classification unit that classifies application usage trends of multiple users into multiple clusters,
A recommendation unit that recommends the setting value of the application to the target user based on the usefulness of the setting value of the application in the cluster to which the usage tendency of the target user belongs and the recommendation condition.
A learning unit that updates at least one of the classification method by the classification unit and the recommendation condition based on the usage tendency, and
Recommendation system with. The recommendation system according to claim 1, wherein the usage tendency includes at least one of the number of times the application is used and the average number of steps. The recommendation system according to claim 1 or 2, wherein the learning unit updates at least one of the classification method and the recommendation condition so as to increase the success probability of the recommendation. The recommendation system according to claim 3, wherein the success of the recommendation includes the condition that the usage tendency of the target user satisfies a predetermined condition after the recommended set value is set by the target user. The recommendation system according to any one of claims 1 to 4, wherein the recommendation unit recommends the set value whose usefulness satisfies the recommendation condition to the target user. Claims 1 to 5 are calculated as the usefulness of the set value increases as the average value of the number of times the application used by the user belonging to the cluster belongs to the set value increases. The recommendation system according to any one of the above. The usefulness of the set value is calculated higher as the average value of the average number of steps of the application corresponding to the set value of the user whose usage tendency belongs to the cluster is smaller. The recommendation system according to any one of the items up to. Any of claims 1 to 7, wherein the usefulness of the set value is calculated higher as the proportion of the user who is setting the set value among the users whose usage tendency belongs to the cluster is larger. The recommendation system described in item 1. The recommendation according to any one of claims 1 to 8, which is calculated higher when the target user who has recommended the set value sets the set value. system. The recommendation system according to any one of claims 1 to 9, wherein the usefulness of the set value is calculated low when the user changes the set value to another set value. The usefulness of the set value is calculated to be high when the number of times the user uses the application corresponding to the set value increases after the user sets the set value. The recommendation system according to any one of up to 10. The usefulness of the set value is claimed from claim 1, which is calculated high when the average number of steps of the application corresponding to the set value of the user decreases after the user sets the set value. Item 2. The recommendation system according to any one of items up to item 11. A classification unit that classifies application usage trends of multiple users into multiple clusters,
A recommendation unit that recommends the setting value of the application to the target user based on the usefulness of the setting value of the application in the cluster to which the usage tendency of the target user belongs and the recommendation condition.
A learning unit that updates at least one of the classification method by the classification unit and the recommendation condition based on the usage tendency, and
Recommendation device equipped with. A classification process that classifies application usage trends of multiple users into multiple clusters,
A recommendation step of recommending the setting value of the application to the target user based on the usefulness of the setting value of the application in the cluster to which the usage tendency of the target user belongs and the recommendation condition.
Based on the usage tendency, a learning step of updating at least one of the classification method by the classification step and the recommendation condition, and
The recommendation method that the computer executes. A classification process that classifies application usage trends of multiple users into multiple clusters,
A recommendation step of recommending the setting value of the application to the target user based on the usefulness of the setting value of the application in the cluster to which the usage tendency of the target user belongs and the recommendation condition.
Based on the usage tendency, a learning step of updating at least one of the classification method by the classification step and the recommendation condition, and
A program that lets your computer run.

Images