KR20180005453A - Apparatus and Method for Analyzing Data using R Association Rule and collaborative filtering - Google Patents

Abstract

The present invention relates to an apparatus and a method for analyzing data by using an R-association rule and collaborative filtering, to detect a user preferred application by applying a collaborative filtering algorithm of R to use time counter data substituted for preference scores necessary to calculate collaborative filtering and a rule generated through an association rule of a programming language R. The apparatus comprises: a collection unit which collects the application list data and application use history data of users; an association rule storage unit which generates association rules by using a programming language R, and stores the rules; a data pre-processing unit which changes the application use history data to data obtained by combining the use time ratio and count ratio; and a selection unit which selects two to ten preferred applications for each user by applying the data obtained by combining the use time ratio and count ratio in the data pre-processing unit to the collaborative filtering algorithm.

Description

[0001] Apparatus and Method for Analyzing Data Using R Association Rule and Collaborative Filtering [

In particular, the present invention relates to an apparatus and method for analyzing data using association rules and collaborative filtering of a programming language R, and more particularly, The present invention relates to an apparatus and method for analyzing data using R association rules and collaborative filtering for detecting preferred applications for each user using R's collaborative filtering algorithm.

With the advent of smartphones and market growth, the handset market is no longer focused on hardware but on software. As customer demands for contents and services increase in the mobile Internet including mobile terminals, hardware - based mobile terminals are no longer appealing to consumers.

In addition, software - centric changes are expanding the application (app: app) ecosystem that can distribute contents and services with diversification of profit model such as contents, service, OS licensing, advertisement and so on to business operators.

Currently, apps installed on smartphones such as iPhone are uploaded and registered in SK Telecom App Store, KT App Store, iPhone App Store and so on.

Looking at the App Store market, the market construction operation is centered on large corporations such as KT, SK, and Apple. The commercial method is limited to the C2C format, and mass market for unspecified users It is a fact that central transactions are being formed.

In addition, since the application type in the App Store is classified into a category type, it is impossible for the category type to perform an ad hoc query by a user, and it is difficult to recommend an application suitable for the user's situation. In addition, according to the other classification method, the application is classified based on the charge rate, and the user selects the application suitable for his / her situation from the recommended application by referring to the ratings and the number of the evaluators.

However, even if there is no guidance on the usefulness or purpose of the application, such a recommended application is not easy to judge which application is suitable for the user due to the scattered configuration, and in the case of the recommended application, There is a problem that it is difficult to be regarded as an application that the user currently needs.

Published Japanese Patent Application No. 10-2008-0026948 (Published Mar. 26, 2008) Japanese Patent Application Laid-Open No. 10-2012-0135396 (published date December 13, 2012)

Therefore, it is an object of the present invention to provide a method and a system for extracting usage time counter data instead of the rules generated through the association rule of the programming language R and the preference score required for the calculation of the collaboration filtering, The present invention provides a data analysis apparatus and method using R association rules and collaborative filtering for detecting a preferred application for each user.

Other objects of the present invention are not limited to the above-mentioned objects, and other objects not mentioned can be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention, there is provided a data analysis apparatus using R association rules and collaboration filtering, comprising: a collection unit for collecting application list data of users and application usage record data; A data preprocessing unit for converting the application use record data into data combining the use time ratio and the count ratio; and a data preprocessing unit for combining the use time ratio and the count ratio in the data preprocessing unit Is applied to a collaborative filtering algorithm and a selection unit for selecting two or more to ten or less preferred applications for each user.

Preferably, when the association rule is created using the programming language R, the association rule storage unit deletes 70 to 80% of users of all users and sets up to search for rules of at least two applications .

Preferably, the data preprocessing unit calculates the application use record data by applying the preference score necessary for the calculation of the collaboration filtering to the use time count data.

Preferably, the use time count data is obtained by adding a use time ratio of each application to the total time of the applications used by the user and a count rate of each application to a total count of the applications used by the user.

According to an aspect of the present invention, there is provided a method of analyzing data using R association rules and collaboration filtering, the method comprising: (A) collecting application list data and application usage record data of users through a collecting unit; ) Generating an association rule indicating a simultaneous detection ratio between the application and the application using the association rule algorithm of the programming language R with the collected application list data through the association rule storage unit, (C) using the application through the data preprocessing unit (D) changing the usage time count data, which is a combination of the usage time ratio and the count ratio, through the selection unit by using a collaborative filtering algorithm of R, And detecting and storing 10 or less preferred applications.

Preferably, the associating rule generated in the step (B) includes, when having an application (lhs) duplicated on the basis of the list of the apps owned by the users, whether or not any other added apps (rhs) .

Preferably, the step (B) includes the steps of, when creating the association rule, deleting, from the calculation for creating the association rule, applications in which 70 to 80% of all the users in the application list data have 70 to 80% And searching for a rule of the abnormal applications.

Preferably, the step (C) changes the user application usage record data to use time count data (rating) required in the calculation of the cooperative filtering, wherein the usage time count data rating And the total count of the applications used by the user is the sum of the count ratios of the respective applications.

The apparatus and method for analyzing data using R association rules and collaborative filtering according to the present invention as described above can generate association rules using user application list data through R association rules and collaborative filtering, It is possible to more accurately recommend the types of applications used by users and applications required for users by detecting a preferred application for each user by using the collaborative filtering algorithm of FIG.

1 is a block diagram showing a configuration of a data analysis apparatus using an R association rule according to an embodiment of the present invention;
2 is a flowchart for explaining a data analysis method using an R association rule according to an embodiment of the present invention.
FIG. 3 is a diagram showing an example of application list data collected by users (rows) in FIG. 2
FIG. 4 is a view showing an example of application usage record data collected by users in FIG. 2
FIG. 5 is a diagram showing an example in which an association rule generated using the association rule algorithm of the programming language R is generated in FIG. 2
Fig. 6 is a view showing an example in which the application use recording data is changed to the use time counter data in which the use time ratio and the count ratio are combined in Fig. 2
FIG. 7 is a diagram showing an example in which five preference application lists are selected for each user

Other objects, features and advantages of the present invention will become apparent from the detailed description of the embodiments with reference to the accompanying drawings.

A preferred embodiment of a data analysis apparatus and method using R association rules and collaborative filtering according to the present invention will be described with reference to the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. It is provided to let you know. Therefore, the embodiments described in the present specification and the configurations shown in the drawings are merely the most preferred embodiments of the present invention and are not intended to represent all of the technical ideas of the present invention. Therefore, various equivalents It should be understood that water and variations may be present.

1 is a block diagram showing a configuration of a data analysis apparatus using an R association rule according to an embodiment of the present invention.

1, a collecting unit 10 for collecting application list data and application use record data of users, an association rule storing unit 20 for generating an association rule using a programming language R and storing the rules, ), A data preprocessing unit (30) for converting the collected application use record data into data combined with a use time ratio and a count ratio, and a data preprocessing unit And a sorting unit 40 for sorting 2 or more to 10 or less preferred applications for each user. The selected user-preferred applications are stored in a storage unit (not shown) in a database.

At this time, when the association rule is created using the programming language R in the classifying unit 20, since almost all the applications such as KakaoTalk are not necessary for calculation, it is necessary to subtract these applications, .

In addition, the analyzer 30 uses the usage time count data instead of the preference score necessary for the calculation of the collaborative filtering on the user application list data. That is, the usage time count data is the sum of the usage time ratio of each application in the total time of the applications used by the user and the total count of the applications used by the user and the count ratio of each application.

The operation of the data analysis apparatus using the R association rule and the collaboration filtering according to the present invention will now be described in detail with reference to the accompanying drawings. The same reference numerals as those in Fig. 1 designate the same members performing the same function.

2 is a flowchart illustrating a data analysis method using an R association rule according to an embodiment of the present invention.

Referring to FIG. 2, first, the application list data of the users and the application usage record data are collected through the collecting unit 10 (S10). FIG. 3 is an embodiment showing application list data collected by users (rows), and FIG. 4 is an embodiment showing application usage record data collected by users.

Next, an association rule indicating a simultaneous detection ratio between the application and the application is generated using the association rule algorithm of the programming language R using the collected application list data through the association rule storage unit 20 (S20). At this time, as shown in FIG. 5, if the generated association rule has an application (lhs) that is duplicated on the basis of the list of the applications that the users have, another application (rhs) to be added is included It is a rule that numerically indicates whether or not it exists.

At this time, when the association rule is created, applications including almost all users such as KakaoTalk, ie, applications including 70% to 80% are unnecessary in calculation for generating association rules. Therefore, It is set to search for the rules of the above applications. The generated association rule simply creates a rule, so no personal characteristics are used.

Then, the application preprocessing unit 30 changes the application use recording data to use time count data obtained by combining the use time ratio and the count ratio (S30). That is, the usage time count data (rating) is the sum of the usage time ratio of each application in the total time of the applications used by the user and the total count of the applications used by the user and the count rate of each application. At this time, FIG. 6 shows an embodiment in which the application use record data is changed to the use time counter data in which the use time ratio and the count ratio are combined in FIG.

Then, the data pre-processing unit 30 detects and stores 10 or fewer preferred applications for each user by using the collaborative filtering algorithm of R on the usage time count data changed by the data preprocessing unit 30 through the selector 40 S50). Here, FIG. 7 shows an embodiment in which five preference application lists are selected for each user.

It is used to analyze the types of applications used by users, or to recommend applications required for users, by using the user-preferred preference applications selected and stored.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. It will be apparent to those skilled in the art that various modifications may be made without departing from the scope of the present invention. Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

Claims (8)

A collection unit for collecting application list data of users and application usage record data;
An association rule storage unit for storing rules by generating an association rule using a programming language R,
A data preprocessing unit for converting the application usage record data into data combining the usage time ratio and the count ratio,
And a selector for applying data obtained by combining the use time ratio and the count ratio in the data preprocessing unit to a collaborative filtering algorithm to select two or more to ten or less preferred applications for each user. And data analysis device using collaborative filtering. The method according to claim 1,
Wherein the classifying unit is configured to delete an application of 70 to 80% of all users when searching for an association rule using a programming language R, and to set a rule to search for a rule of at least two applications. Data analysis device using rules and collaborative filtering. The method according to claim 1,
Wherein the data preprocessor preprocesses the application use record data with usage time count data. The method of claim 3,
Wherein the usage time count data is obtained by summing a use time ratio of each application in the total time of the applications used by the user and a count ratio of each application to a total count of applications used by the user. Analysis device. (A) collecting application list data of the users and application usage record data through the collection unit;
(B) generating an association rule indicating the simultaneous detection ratio between the application and the application using the association rule algorithm of the programming language R with the application list data through the association rule storage unit;
(C) changing the application use record data to data combined with the use time ratio and the count ratio through a data preprocessing unit;
(D) a step of detecting data of two or more to ten or less preferred applications for each user by using a collaborative filtering algorithm of R and combining the use time ratio and the count ratio in the data preprocessing unit through the selector And the R-association rule and the collaborative filtering. 6. The method of claim 5,
If the association rule generated in the step (B) has a duplicate application (lhs) based on the list of applications owned by the users, it is possible to numerically indicate whether any other app (rhs) And a data analysis method using collaborative filtering. 6. The method of claim 5, wherein step (B)
When creating an association rule, the applications having 70 to 80% of all users in the application list data are deleted from the calculation for creating the association rule,
And setting a rule of at least two of the application list data to search for a rule of at least two of the application list data. 6. The method of claim 5,
In the step (C), the user application usage record data is changed to the usage time count data required in the calculation of the cooperative filtering. In this case, the usage time count data (rating) And a counting ratio of each application to a total count of the applications used by the user, and a data analysis method using collaborative filtering.

Images