JP5481289B2 - Server and method for recommending applications to users - Google Patents

Description

  The present invention relates to a server and method for recommending an application to a user.

  Recent communication terminals not only perform communication, but also can download and execute various applications from a server. Specifically, the communication terminal is a mobile phone, an information terminal, a smartphone, a notebook personal computer, a laptop computer, a desktop computer, or the like, but is not limited thereto. Since there are an extremely large number of applications that can be downloaded to the communication terminal, the application provider (provider) needs to recommend the application to be provided to the user. Downloading an application to a communication terminal is described in Patent Document 1, for example.

  One method of determining an application recommended by a provider to a user is to determine similarity between users and preferentially recommend an application downloaded by a similar user. For example, user A has downloaded applications 1, 2, 3, 4 to his / her communication terminal, user B has downloaded applications 1, 2, P, Q to his / her communication terminal, and user C himself / herself Suppose that applications V and W have been downloaded to the communication terminal. When determining an application recommended by the provider to the user A, first, the similarity between the users is determined. User B owns the same applications 1 and 2 as user A, but user C does not own any of the same applications as user A. Therefore, it can be seen that the users A and B have similar preference trends (preferences), and both are similar, but the users A and C are not similar. In this way, it is understood that the provider only recommends to the user A the applications P and Q not owned by the user A among the applications owned by the user B.

JP 2009-301472 A

  By the way, with the spread of communication terminals, one user may have a plurality of communication terminals. In this case, when there is one user account for a plurality of communication terminals, the provider recommends the same application to a user having a plurality of communication terminals without distinguishing the plurality of communication terminals. Will do. Usually, a plurality of terminals owned by the same user have different product specifications (CPU processing speed, display screen size, etc.), and operating environments are different. For such a plurality of communication terminals, there is a concern that even if an application that the user does not own is simply recommended, it is not necessarily meaningful for the user. This is because an application may work well on one communication terminal but not on another communication terminal. As described above, when an application is recommended to a user, the communication terminal used by the user, that is, the operating environment is an extremely important problem. In this respect, the situation is different from the case where mere content that is not an application is recommended to the user. This is because the user's operating environment is not as important when recommending content as it is with applications.

  From the viewpoint of considering such differences between communication terminals, it is conceivable to assign different user accounts to a plurality of communication terminals owned by the same user. However, if the user account is different, management is performed in the same manner as when the users of the plurality of communication terminals are different. In the first place, the user's preference tendency can be properly determined by considering the entire application downloaded to a plurality of communication terminals owned by the user. It is not easy to appropriately read the user's preference tendency from only some of the applications downloaded to each of the plurality of communication terminals.

  Therefore, in the conventional method, when the user has a plurality of communication terminals, there is a concern that an application suitable for each communication terminal cannot be recommended appropriately.

  An object of the present invention is to enable a server to recommend an application suitable for each communication terminal to a user when the user has a plurality of communication terminals in a system in which the server communicates with the user's communication terminal via a communication network. It is to be.

A server according to an aspect of the present invention includes:
A server in a communication system including at least a first user having a plurality of communication terminals and a second user having a terminal having the same specification as any of the plurality of communication terminals,
A receiving unit that receives usage history information indicating the usage status of an application from each user of the communication system;
A tabulation unit that tabulates the usage history information and creates a master list including candidate applications recommended for the first user;
According to the usage trend of a certain communication terminal used by the first user or the usage trend of the terminal used by the second user, for each candidate application included in the master list, the first user's A recommended app determining unit that determines priorities recommended for each communication terminal;
A transmission unit that transmits a recommendation signal to the first user so as to present the candidate to the first user according to the priority.

  According to an aspect of the present invention, in a system in which a server communicates with a user's communication terminal via a communication network, when the user has a plurality of communication terminals, the server recommends an application suitable for each communication terminal to the user. it can.

The figure which shows the communication system used in an Example. The functional block diagram of a communication terminal. The functional block diagram of a delivery server. The operation | movement conceptual diagram in an Example. The flowchart of the operation example used in an Example. The flowchart of the operation example at the time of calculating similarity degree sim (A, B). The flowchart in the case of weighting based on the usage tendency of the whole user. The flowchart in the case of performing weighting based on the usage tendency (category) of the recommendation target user. The figure which shows the example of calculation of the utilization factor of the application of each category in each communication terminal. The flowchart in the case of weighting based on the usage tendency (feature vector) of the recommendation target user.

  Examples will be described from the following viewpoints.

1. System 2. 2. Communication terminal Server 4. Example of Operation 4.1 Outline of Operation 4.2 Operation Flow 4.3 Specific Example of Weighting 4.3.1 Weighting Based on Usage Trend of Entire User 4.3.2 Weighting Based on Usage Trend (Category) of Recommended User 4 3.3 Weighting based on usage trends (feature vectors) of recommended users

<1. System>
FIG. 1 shows a communication system 1 used in the embodiment. The communication system 1 includes a communication terminal 10, a mobile communication network 20, a distribution server 30 and an application server 40. The communication terminal 10 communicates with the distribution server 30 via the mobile communication network 20. The distribution server 30 communicates with the application server 40 as necessary. The application server 40 has an application that can be provided by the provider, and provides the application in response to a request from the distribution server 30.

<2. Communication terminal>
FIG. 2 shows a functional block diagram of the communication terminal 10. FIG. 2 shows various functional elements included in the communication terminal 10 that are particularly relevant to the present embodiment. The communication terminal may be any appropriate communication terminal that can download and execute an application from the distribution server. Specifically, the communication terminal is a mobile phone, an information terminal, a smartphone, a notebook personal computer, a laptop computer, a desktop computer, or the like, but is not limited thereto. The illustrated communication terminal 10 includes at least a reception unit 101, a display unit 102, a usage history acquisition unit 103, and a transmission unit 104.

  The receiving unit 101 receives an application and other signals from the distribution server via the mobile communication network. Other signals include, for example, a recommendation signal that prompts the user to download some application.

  The display unit 102 functions as a visual user interface. Specifically, the display unit 102 is a liquid crystal display (LCD), a plasma display, an organic EL panel, or the like, but is not limited thereto. The display unit 102 in the present embodiment realizes a touch screen having a touch-sensitive transparent panel, and can control the display of the screen by the movement of the user's finger. The display unit 102 can display characters, graphics, symbols, still images, moving images, and the like. For example, an icon representing an application is displayed on the screen.

  Not only a visual user interface such as the display unit 102 but also other user interfaces may be provided in the communication terminal 10. For example, a keypad, a control panel, a keyboard, a touchpad, a mouse, a trackball, a microphone, a speaker, and the like may be provided.

  The usage history acquisition unit 103 creates usage history information indicating application download, activation, deletion, and the like. For example, when a certain application is downloaded, the application name and the download date / time are recorded at least in the usage history information. Furthermore, when an application is activated, the name of the target application, the date and time of activation, the period of use, the number of times of activation, and the like are recorded in at least usage history information. When an application is deleted, the target application name and the date and time of deletion are recorded at least in the usage history information. The usage history information may include information on the location where the application is used, in addition to the application name, date and time, usage time, and the like.

  The transmission unit 104 transmits a signal to the distribution server via the mobile communication network. The transmitted signal includes, for example, the above usage history information. The usage history information may be reported to the distribution server at a certain frequency, or may be reported in response to a request from the user, the distribution server, or the like. The frequency when reporting the usage history information at a certain frequency may be any appropriate frequency such as time, day, week, month, and the like. From the viewpoint of storing the latest information in the distribution server and reducing the amount of usage history information stored in the communication terminal, it is desirable to report at a high frequency. From the viewpoint of effective use of network resources, it is desirable to report as necessary.

<3. Server>
FIG. 3 shows a functional block diagram of the distribution server 30. FIG. 3 shows various functional elements provided in the distribution server 30 that are particularly relevant to the present embodiment. The distribution server 30 searches for an application stored in the application server 40 and provides the application to the user of the communication terminal via the mobile communication network. The illustrated distribution server 30 includes at least a reception unit 301, a storage unit 302, a usage history totaling unit 303, a user similarity calculation unit 304, a recommended application determination unit 305, a metadata processing unit 306, and a transmission unit 307.

  The receiving unit 301 receives a signal from a communication terminal via a mobile communication network. Other signals include, for example, usage history information. The usage history information is received from each of a plurality of communication terminals connected to the distribution server 30.

  The storage unit 302 stores parameters, calculation results, and information used in the distribution server. In particular, usage history information received from each user by the receiving unit 301 is stored.

  Based on the usage history information received from each user, the usage history totaling unit 303 creates a master list in which candidates for applications recommended for specific users are listed, and stores the master list in the storage unit 302. In general, a master list is a list that lists applications that a particular user (eg, user A) has not downloaded to any of their communication terminals. This list preferentially includes an application downloaded by a user (for example, user B) similar to a specific user (user A) to his or her (user B) communication terminal.

  The user similarity calculation unit 304 calculates the degree of similarity (similarity) between users. The similarity between users may be calculated based on any appropriate standard. For example, it may be calculated from the viewpoint of whether or not the application that the user is downloading to his / her communication terminal is common among the users. For example, the greater the number of applications downloaded by both users A and B, the greater the similarity between them (similar). Alternatively, the similarity may be calculated based on a tendency to use the application (for example, usage frequency). A specific method for calculating the similarity will be described later.

  The recommended application determining unit 305 selects an application recommended for a specific user (user A) having a plurality of communication terminals (UE_A1, UE_A2) from the master list. The selection is based on the usage tendency of another user X having a communication terminal similar to the communication terminal (UE_A1) of user A and the usage tendency of another user Y having a communication terminal similar to the communication terminal of user A (UE_A2). May be performed on the basis. Alternatively, the selection from the master list may be performed based on the usage tendency of the communication terminal (UE_A1) of the user A or the usage tendency of another communication terminal (UE_A2) of the user A. Note that the communication terminal used by the user X is not necessarily the same as the communication terminal (UE_A1) of the user A. A user of a communication terminal that conforms to the same product specifications as the communication terminal (UE_A1) of user A corresponds to user X. Similarly, the communication terminal used by the user Y may not necessarily be the same as the communication terminal (UE_A2) of the user A. A user of a communication terminal that conforms to the same product specifications as the communication terminal (UE_A2) of user A corresponds to user Y. Product specifications are items that affect the operating environment of applications such as CPU processing speed, display screen size, operating system, and the like.

  How the recommended application determining unit 305 selects a recommended application from the master list candidates will be described later.

  As described above, the metadata processing unit 306 generates a feature vector from the metadata of the application. For example, in the case of a transfer guidance application, words such as transfer, station, and timetable are extracted from the metadata as characteristic words, and a feature vector is created from these words.

  The transmission unit 307 creates a recommendation signal that recommends downloading of an application decided to be recommended to the user, and transmits the recommendation signal to the user.

<4. Example of operation>
<4.1 Outline of operation>
FIG. 4 conceptually shows the operation in this embodiment. Three users A, B, and C are shown, but there are actually many users. User A has three communication terminals, and has an information terminal (UE_A1) such as a smartphone, another information terminal (UE_A2), and a television (UE_A3). User B has two communication terminals, and has an information terminal (UE_A2) and a car navigation (UE_B1). User C has three communication terminals, and has an information terminal (UE_A1), a television (UE_A3), and a car navigation (UE_B1).

  In the present embodiment, usage histories of a plurality of communication terminals are aggregated for each user and reported appropriately to a distribution server (not shown). That is, there is one user account for a plurality of communication terminals. However, as will be described later, a weighting factor corresponding to each terminal is considered. The distribution server (not shown in FIG. 4) calculates the similarity between users. In the illustrated example, the similarity between the users A and B is γ, the similarity between the users B and C is β, and the similarity between the users A and C is α. The distribution server creates, for example, a master list for user A based on the usage history of each user.

  The master list is a list of applications that user A has not downloaded. For example, user A has downloaded applications 1, 2, 3, 4 to his / her communication terminals (UE_A1, UE_A2, UE_A3), and user B has applications 1, 2, a to his / her communication terminals (UE_B1, UE_A2). , B, and c, and the user C has downloaded the applications V and W to his / her communication terminals (UE_A1, UE_A3, and UE_B1). User B owns the same applications 1 and 2 as user A, but user C does not own any of the same applications as user A. Therefore, from the viewpoint of the similarity of the applications, it can be seen that the users A and B have similar preference tendencies (preferences), but the users A and C are not similar. In the illustrated example, users A and B have an information terminal (UE_A1) and a television (UE_A3). Users A and C have information terminals (UE_A2). Therefore, it can be seen that the user B is similar to the user A from the viewpoint of the similarity of the owned communication terminals, but the user C is not similar to the user B. From this point of view, the master list for user A can be determined to include, for example, applications a, b, and c.

  When the distribution server recommends an application included in the master list to the user A, the distribution server recommends an application candidate suitable for the specific communication terminal of the user A according to the usage tendency of the application of the user A or another user.

<4.2 Operation flow>
FIG. 5 shows a flow chart showing how the distribution server recommends the application candidates in the master list to the user.

  In step S51, the distribution server determines the rate r ′ (A, a) of a specific user A for an unevaluated candidate a, the rate r (j, a) of another user, and the similarity sim (A, j) between users. ) (Similarity between user A and user j). An unevaluated candidate is a candidate for which a non-zero rate has not yet been estimated among application candidates included in the master list for user A. The distribution server first calculates the rating vector and similarity of each user.

  FIG. 6 shows a flowchart for calculating the similarity sim (A, B). In step S61, the user similarity calculation unit 304 of the distribution server collects the usage history of each user's application from the storage unit 302 for each terminal. The usage history information is appropriately received by the distribution server and stored in the storage unit 302 as needed. For example, since the user A has a plurality of communication terminals (UE_A1, UE_A3, UE_B1), a usage history for the communication terminal (UE_A1), a usage history for the communication terminal (UE_A2), and a usage history for the communication terminal (UE_A3) are collected. Is done.

  In step S62, the usage history totaling unit 303 of the distribution server totals the usage history of the application for each terminal into the usage history for each user. For example, the usage history of the communication terminal (UE_A1) of the user A indicates that the application p has been started six times, and the usage history of the communication terminal (UE_A2) of the user A indicates that the application p has been started four times. Suppose. In this case, these usage histories are collected into information indicating that the user A has activated the application p ten times.

In step S63, a rating vector is calculated for each user. The rating vector represented by “vector r” is an M-dimensional vector. M represents the number of applications present in the market. Alternatively, M is the number of applications that the provider can provide. The M components of user A's rating vector “vector r _A ” indicate how much the application corresponding to each component is being used, and is referred to as “rate”. For example, in the above example, the user A has used the application p ten times in total. Accordingly, the component (rate) corresponding to the application p among the M components of the rating vector “vector r _A ” of the user A has a value of “10” (r (A, p) = 10). As an example, assume that there are four market applications, p, q, s, and t (M = 4). User A has these applications p and is activated 10 times. Assume that no other applications have been downloaded. In this case, the rating vector “vector r _A ” of user A is
(R (A, p), r (A, q), r (A, s), r (A, t)) = (10, 0, 0, 0)
It can be expressed as Alternatively, a value normalized or normalized with some value, not the number of times of use itself, may be set as the component value.

  In step S64, the similarity between users is calculated according to the following formula (1) using the rating vector of each user.

ユーザＡは、マスターリスト中のアプリケーションの候補ａを未だダウンロードしていない。したがって、ユーザＡのレーティングベクトルの候補ａに対するレートｒ（Ａ，ａ）は、０の値をとることになる。本実施例では、このレートｒ（Ａ，ａ）＝０をそのまま使用するのではなく、実際のレートｒ（Ａ，ａ）＝０の代わりに、推定されたレートｒ'（Ａ，ａ）を使用する。推定されたレート（推定レート）ｒ'（Ａ，ａ）は、以下の数式（２）にしたがって算出される。

User A has not yet downloaded the application candidate a in the master list. Accordingly, the rate r (A, a) for the candidate a of the rating vector of user A takes a value of 0. In the present embodiment, this rate r (A, a) = 0 is not used as it is, but the estimated rate r ′ (A, a) is used instead of the actual rate r (A, a) = 0. use. The estimated rate (estimated rate) r ′ (A, a) is calculated according to the following equation (2).

上記の数式を参照すると、ユーザＡと類似しているユーザｊが、アプリケーションａを何回も利用していた場合、推定レートｒ'（Ａ，ａ）は大きくなることが分かる。マスターリストに含まれている全ての候補ａ、ｂ、ｃ、...について、推定レートｒ'（Ａ，ａ）、ｒ'（Ａ，ｂ）、ｒ'（Ａ，ｃ）、...が算出される。

Referring to the above mathematical expression, it can be seen that the estimated rate r ′ (A, a) increases when the user j similar to the user A has used the application a many times. For all candidates a, b, c,... Included in the master list, the estimated rates r ′ (A, a), r ′ (A, b), r ′ (A, c),. Is calculated.

  In step S52 of FIG. 5, the estimated rates r ′ (A, a), r ′ (A, b), r ′ (A, c),... Calculated in step S51, and weighting factors described later are used. To determine the priority of the application candidates a, b, c,... As will be described later, the weighting factor and the priority are greatly different from the conventional application recommendation method in that the weighting factor and the priority are set to values suitable for each distribution destination terminal.

  In step S53, candidate applications are presented to the user in descending order of priority. Thereby, the application suitable for each communication terminal is displayed in a meaningful order on the display screen of each of the plurality of communication terminals (UE_A1, UE_A3, UE_B1) of the user A, and the user can benefit from the excellent recommendation function. Can receive.

<4.3 Specific examples of weighting>
As described above, in step S52 in the flow of FIG. 5, the estimated rates r ′ (A, a), r ′ (A, b), r ′ (A, c),. The priority of the application candidates a, b, c,. For convenience of explanation, as in the example shown in FIG. 4, the user A has three communication terminals, an information terminal (UE_A1) such as a smartphone, another information terminal (UE_A2), and a television (UE_A3). ). User B has two communication terminals, and has an information terminal (UE_A2) and a car navigation (UE_B1). The user C has three communication terminals, and has an information terminal (UE_A1), a television (UE_A3), and a car navigation (UE_B1). In such a situation, how to determine the weighting factor in step S52 of FIG. 5 will be described.

<4.3.1 Weighting based on usage tendency of the entire user>
The first method for determining the weighting factor is to use the usage trend of the entire user. FIG. 7 shows a flowchart for performing such weighting. In step S <b> 71, the recommended application determination unit 305 of the distribution server extracts the usage history information of the entire user from the storage unit 302 for the candidate a included in the master list.

  In step S72, the extracted usage history information is tabulated for each terminal.

  In step S <b> 73, the ratio of users who use candidate applications among users having similar terminals is calculated. For example, the number of users or the percentage of users who have downloaded the application of candidate a among the users having the same terminal as the communication terminal (UE_A1) (utilization rate 1), and the same terminal as the communication terminal (UE_A2) The number of users or the percentage of users who have downloaded (number of users 2) is calculated.

  For example, it is assumed that there are 100 users X having terminals having the same specifications as the communication terminal (UE_A1) of user A, and application candidates a included in the master list have been downloaded to 70 users X. In this case, the utilization factor 1 for the candidate a is 70/100 = 0.7. Furthermore, for another communication terminal (UE_A2) of user A, the usage tendency of another user Y having a terminal with the same specifications as this communication terminal (UE_A2) is used. For example, it is assumed that there are 50 users Y and the candidate a for the application included in the master list has been downloaded only to two users Y out of 50. In this case, the utilization rate 2 for the candidate a is 2/50 = 0.04.

  In step S74, the usage rate calculated in step S72 is set as a weighting factor for each of the plurality of communication terminals of the target user A. For example, the weighting factor for user A's communication terminal (UE_A1) is set to utilization factor 1 = 0.7, and the weighting factor for user A's communication terminal (UE_A2) is set to utilization factor 2 = 0.04. .

  In this way, it is possible to appropriately determine how strongly an application candidate is recommended to the user A based on a usage tendency of a user who has a terminal having the same specification as the communication terminal of the user A.

  Thereafter, the priority of the candidates a, b, c,... In the master list is determined by combining this weighting factor and the rate estimated in step S51 of FIG. As an example, the priority may be determined by multiplying a weighting factor and a rate. For example, the candidate a is weighted to the communication terminal (UE_A1) of the user A as r ′ (A, a) × (utilization rate 1). In addition, the candidate a is weighted to the communication terminal (UE_A2) of the user A as r ′ (A, a) × (utilization rate 2). Based on these weighting results, the priority of each candidate is determined.

  According to this weighting method, an application suitable for the communication terminal can be recommended. For example, a moving image application can be recommended for a communication terminal having a large screen display unit, while a transfer guidance application can be recommended for a mobile phone.

<4.3.2 Weighting based on usage trend (category) of recommended users>
A second method for determining the weighting factor is to use the usage tendency of the recommendation target user (user A).

  FIG. 8 shows a flowchart for performing such weighting. In step S <b> 81, the recommended application determining unit 305 of the distribution server extracts the usage history information of the user A from the storage unit 302.

  In step S82, the usage history information of the user A is aggregated for each application category. A category is a tool, a transfer guide, a game, or the like, and is an information item for classifying a plurality of applications.

  In step S83, the usage rate for each terminal is calculated from the counting result in step S82. For example, the category of 50 applications downloaded by user A to three communication terminals (UE_A1, UE_A2, UE_A3) is category 1 (20, 40%), category 2 (15, 30%), Assume that category 3 (10, 20%) and category 4 (5, 10%). It is assumed that ten category 1 applications have been downloaded to the communication terminal (UE_A1), six to the communication terminal (UE_A2), and four to the communication terminal (UE_A3). In this case, the utilization rate of the category 1 application in the communication terminal (UE_A1) is 10/20 = 0.5. Similarly, the usage rate of the application of category 1 in the communication terminal (UE_A2) is 6/20 = 0.3. The usage rate of the category 1 application in the communication terminal (UE_A3) is 4/20 = 0.2.

  FIG. 9 shows a calculation example of the utilization rate of each category of application in each communication terminal. The above considerations for category 1 are summarized in the category 1 row of FIG. In the figure, the numbers in parentheses in the percentage column indicate how many of the 50 applications. The numbers in parentheses in each column of the communication terminals (UE_A1, UE_A2, UE_A3) indicate the utilization rate.

  Thus, in step S83, the utilization factor for each communication terminal is calculated for each category.

  In step S84, the usage rate calculated in step S83 is set as a weighting factor for each of the plurality of communication terminals of the target user A. For example, when the category of the application candidate a is category 2, the weighting factor for the communication terminal (UE_A1) is 10/15 = 0.667, and the weighting factor for the communication terminal (UE_A2) is 5/15. = 0.334, and the weighting factor for the communication terminal (UE_A3) is 0.

  In this way, it is possible to appropriately determine how strongly an application candidate is recommended to the user A based on the usage tendency of each communication terminal of the user A.

  Thereafter, the priority of the candidates a, b, c,... In the master list is determined by combining this weighting factor and the rate estimated in step S51 of FIG. As an example, the priority may be determined by multiplying a weighting factor and a rate. For example, the category 2 candidates are weighted to the communication terminal (UE_A1) of user A as r ′ (A, a) × (0.667). The category 2 candidates are weighted to user A's communication terminal (UE_A2) as r ′ (A, a) × (0.334). Further, the category 2 candidates are weighted such that r ′ (A, a) × (0) = 0 for the communication terminal (UE_A3) of the user A. Based on these weighting results, the priority of each candidate is determined.

  According to this weighting method, it is possible to recommend an application in accordance with the usage of the communication terminal of each user. For example, while recommending a game application to a communication terminal that frequently uses a puzzle game, it is possible to recommend a transfer guidance application to a communication terminal that frequently uses a map search application. .

<4.3.3 Weighting based on usage tendency (feature vector) of recommended users>
Similar to the second method, the third method for determining the weighting factor uses the usage tendency of the user to be recommended (user A), but differs in that a feature vector is used instead of a category.

  FIG. 10 shows a flowchart for performing such weighting. In step S101, the metadata processing unit 306 of the distribution server discriminates and extracts characteristic words that are characteristic words from the metadata of all applications that can be provided by the provider, and determines a total of n characteristic words. . The extraction of the feature words may be performed by any appropriate method in the technical field. As an example, it may be performed by TFIDF. Note that the metadata in this case is information that characterizes the application, such as an application name and an application description.

  In step S102, the metadata processing unit 306 generates a feature vector for each of the application candidates included in the master list for the user A. The feature vector is an n-dimensional vector, and is generally generated by determining whether or not n feature words determined in step S101 are included.

  In step S103, the metadata processing unit 306 generates a feature vector for each communication terminal of the user A based on the usage history information of the user A. For example, it is assumed that the communication terminal (UE_A1) of the user A has downloaded the map search application (Ap1) and the music appreciation application (Ap2). In this case, one feature vector is generated using the metadata of each of the two applications Ap1 and Ap2. Similarly, feature vectors are also generated for other communication terminals of user A.

  In step S104, the similarity between the feature vector of the application candidate generated in step S102 and the feature vector of each terminal generated in step S103 is calculated, and the similarity is determined as a weighting factor.

  In this way, it is possible to appropriately determine how strongly an application candidate is recommended to the user A based on the usage tendency of each user A's communication terminal.

  Thereafter, the priority of the candidates a, b, c,... In the master list is determined by combining this weighting factor and the rate estimated in step S51 of FIG. As an example, the priority may be determined by multiplying a weighting factor and a rate.

  Although the present invention has been described with reference to particular embodiments, they are merely exemplary and those skilled in the art will appreciate various variations, modifications, alternatives, substitutions, and the like. For example, the present invention may be applied to any appropriate mobile communication system that recommends applications to users. For example, the present invention includes W-CDMA system, HSDPA / HSUPA W-CDMA system, LTE system, LTE-Advanced system, IMT-Advanced system, WiMAX, Wi-Fi system, etc. May be applied. Although specific mathematical formulas have been described to facilitate understanding of the invention, these mathematical formulas are merely examples, unless otherwise specified, and any appropriate mathematical formula may be used. The classification of the examples or items is not essential to the present invention, and the items described in two or more items may be used in combination as necessary, and the items described in one item may be combined with other items. It may be applied (as long as it is not inconsistent) to the matters described. For convenience of explanation, an apparatus according to an embodiment of the present invention has been described using a functional block diagram, but such an apparatus may be realized by hardware, software, or a combination thereof. The software is available on random access memory (RAM), flash memory, read only memory (ROM), EPROM, EEPROM, registers, hard disk (HDD), removable disk, CD-ROM, database, server and any other suitable storage medium May be. The present invention is not limited to the above embodiments, and various modifications, modifications, alternatives, substitutions, and the like are included in the present invention without departing from the spirit of the present invention.

DESCRIPTION OF SYMBOLS 1 Communication system 10 Communication terminal 101 Reception part 102 Display part 103 Usage history acquisition part 104 Transmission part 20 Mobile communication network 30 Distribution server 301 Reception part 302 Storage part 303 Usage history total part 304 Inter-user similarity calculation part 305 Recommended application determination part 306 Metadata processing unit 307 transmission unit 40 application server

Claims (2)

A server in a communication system,
A receiving unit that receives usage history information indicating the usage status of an application from each user of the communication system;
A counting unit that creates a master list comprising a candidate application aggregates the use history information, recommended to the first user,
A metadata processing unit that creates a feature vector for the application by extracting characteristic words from the application information;
Therefore the use trend of the first users to use communication terminals, for each candidate application included in the master list, met the recommended application determination unit determines the priority to be recommended to the first user Depending on the similarity between the feature vector created from the candidate information included in the master list and the feature vector created from the application information downloaded to the communication terminal of the first user, A recommended application determining unit that determines the priority ;
A transmission unit that transmits a recommendation signal to the first user so as to present the candidate application to the first user according to the priority. A recommendation method for the server to perform the communication system,
From the user each of said communication system, receiving the usage history information indicating a usage of the application,
And creating a master list including application candidates that aggregates the use history information, recommended to the first user,
Creating a feature vector for the application by extracting characteristic words from the application information;
Said first user thus can be that utilizing trend of communication terminals used for candidate each application included in the master list, and determining the priority to be recommended to the first user, the According to the similarity between the feature vector created from the candidate information included in the master list and the feature vector created from the application information downloaded to the communication terminal of the first user, the priority Determine the steps , and
The application of the candidate according to the priority to present to the first user, the recommendation method and a step of transmitting a recommendation signal to said first user.

Images