JP4308683B2 - User activity history visualization / analysis method, user activity history visualization / analysis device, and program - Google Patents

Description

  The present invention relates to technology for visualizing and analyzing user activity histories, and in particular, comprehending a large amount of activity histories of users who use various services via an information terminal after that activity to grasp the attitude. In addition, the present invention relates to a technique for visualizing and analyzing in a batch so that details can be traced.

  As services operating on computers and computer networks have become widespread, it has become possible to obtain abundant activity histories on the terminals of service users (users).

  For example, it is possible to automatically and detailedly acquire user information collection, decision making, and history of thought processes that were difficult to record with conventional face-to-face and paper-based services, It is expected to be used in fields such as education.

  In general, a statistical method is often applied as a method for visualizing and analyzing the collected history. For example, in the standard WWW server Apache (http://httpd.apache.org) Analog (http://www.analog.cx/), which is a report tool for reporting the access status, simple access In addition to the number, the bias of access in the WWW site is aggregated and visualized in relation to the date and time and the user, and the usage status of the site and the characteristics of the user are analyzed (for example, non-patented Reference 1).

Alternatively, the learning management system (LMS), which is an online standardized learning management system, can generate and present a learning progress report from a standardized history format (for example, non-patented). References 2 and 3.)
Stephen Turner, “Analog-The most popular logfile analyzer in the world”, [online], last updated November 30, 2003, [January 14, 2004 search], Internet <URL: http: // www. analog. cx /> NTT-X Co., Ltd., “XcalatII-LMS function”, [searched on January 14, 2004], Internet <URL: http: // e-cube. ne. jp / Solution / Products / xcalat2. sht> Digital Forest Co., Ltd., “Visionalist WEB Analysis-Product Introduction”, [searched on January 14, 2004], Internet <URL: http: // www. visionalist. com / web / seihin. html>

  As described above, detailed and abundant activity history can be obtained automatically and easily, while analysis of a large amount of information and feedback to analysts, for example, visualization methods are more difficult. It has become a thing.

  For example, in the conventional statistical method, a constant result can be calculated regardless of the amount of history, but the result is comprehensive and elements such as characteristics and styles collected for each user are lost. This is not necessarily sufficient for marketing purposes, such as extracting behavioral trends and preferences of buyers in online shopping, and is also appropriate as feedback for individuals such as mentoring (individual learning guidance) in online learning services. I can't say that.

  In recent years, a new analysis method for a series of meaningful activity histories called click stream analysis has been introduced and used from the marketing aspect to deal with such problems (eg: http://www.visionalist.com/web/index.html).

  However, the click stream analysis is based on detailed operations depending on operations unique to each site (service), links, and the like, and the difficulty of individual user analysis increases as the amount of history increases. In addition, it is basically limited to user actions within the same site (service), and creating a place where users can use various services in parallel or switching as in recent years depends on each service. Since the entire history is included, the analysis becomes complicated, making it difficult for the analyst to grasp.

  The present invention has been made in view of the above points, and a list of complex and abundant activity histories of users who use various services via an information terminal is listed after the activity, and the details are understood. Visualization of user activity histories that can be grouped from the viewpoint of activity trends that are different from hobbies and preferences, by calculating feature vectors that represent activity trends in a batch The purpose is to provide analysis technology.

In order to solve the above problems, the invention of claim 1 is an information terminal in which a user performs a service request operation, a service providing server that receives a user request from the information terminal and provides a service, and the information terminal. And a user activity history visualization / analysis device that visualizes a user activity history in a service providing server, wherein the information terminal adds time information to a user operation content. A step of recording as a user activity history, a step of the service providing server adding time information to the user request contents and recording as a user activity history, and a user activity history visualization / analysis device including the information terminal and Acquiring the user activity history recorded in the service providing server; A step of attaching a classification ID to the user activity history according to the operation content and the request content, a step of sorting the acquired user activity history based on time information, and a user activity history for the sorted user activity history Calculating and adding an elapsed time between, creating a time-series activity transition status list for the operation content and request content, and visualizing the time-series activity transition status list , The context of the user's operation content and request content is reflected in the value of the classification ID, and the value of the classification ID is also used as a key for the sorting.

According to a second aspect of the present invention, there is provided an information terminal in which a user performs a service request operation, a service providing server that accepts a user request from the information terminal and provides a service, and the information terminal and the service providing server. A user activity history visualization / analysis method in a service providing system having a user activity history visualization / analysis device that visualizes a user's activity history, wherein the information terminal attaches time information to a user's operation content and the user activity A step of recording as a history, a step in which the service providing server adds time information to the user's request content and recording as a user activity history, and a recording and recording of the user activity history in the information terminal and the service providing server. Acquiring the acquired user activity history, and acquiring the acquired user activity history The step of attaching a classification ID according to the operation content and the request content, the step of sorting the acquired user activity history based on time information, and the elapsed time between user activity histories for the sorted user activity history Calculating and adding, a step of creating a time-series activity transition status list for the operation content and the request content, and a total activity time for each classification ID from the created time-series activity transition status list The method includes a step of calculating a ratio value of the occupied activity time and a step of visualizing a result of the ratio value calculation.

The invention according to claim 3 is characterized in that, in claim 2, the activity value of the user having the same dimension as the number of the category IDs is indicated from the ratio value of the activity time to the total activity hours for each category ID. The method includes a step of obtaining a vector, and a step of extracting users having similar activity tendencies using the feature vector calculated for each user.

According to a fourth aspect of the present invention, an information terminal on which a user performs a service request operation records a user activity history recorded with time information added to a user operation content, and receives a user request from the information terminal. A service activity server for providing a service receives a user activity history recorded with time information added to a user request content, and visualizes the user activity history. Means for acquiring a user activity history recorded in a terminal and a service providing server; means for attaching a classification ID to the acquired user activity history according to an operation content and a request content; and the acquired user activity history. A means for sorting based on time information, and an elapsed time between user activity histories for the sorted user activity histories are calculated and attached. Means for creating a time-series activity transition status list for the operation content and the request content, and means for visualizing the time-series activity transition status list. The context is reflected in the value of the classification ID, and the value of the classification ID is also used as a key for the sorting.

According to a fifth aspect of the present invention, a user activity history recorded by an information terminal on which a user performs a service request operation with time information added to the user's operation content, and a user request from the information terminal are received. A service activity server for providing a service receives a user activity history recorded with time information added to a user request content, and visualizes the user activity history. Means for acquiring a user activity history recorded in a terminal and a service providing server; means for attaching a classification ID to the acquired user activity history according to an operation content and a request content; and the acquired user activity history. A means for sorting based on time information, and an elapsed time between user activity histories for the sorted user activity histories are calculated and attached. A means for creating a time-series activity transition status list for the operation content and the request content, and a ratio of activity time to the total activity time for each classification ID from the created time-series activity transition status list And means for calculating a value and means for visualizing the result of the ratio value calculation.

The invention according to claim 6 is characterized in that, in claim 5, the activity tendency of the user of the same dimension as the number of the category IDs is determined from the ratio value of the activity hours to the total activity hours for each category ID. It has a means to acquire a vector, and a means to extract users with a similar activity tendency using the feature vector computed for every user.

According to a seventh aspect of the invention, the user activity history visualization / analysis method or the user activity history visualization / analysis apparatus according to any one of the first to sixth aspects is described by a computer program. It is characterized by being executable.

In claim 1.2.4.5 , the service providing server may be one or more. The service providing server and the information terminal can be connected within the same terminal or via a network.

  When the classification ID is attached according to the operation content and the request content, a different classification ID can be attached to each service providing server, each service providing server function, each information terminal, and each information terminal function.

  When visualizing the time series activity transition list, a rectangle having a width corresponding to the elapsed time corresponding to each history is drawn in a color set for each classification ID corresponding to the history, and this is repeated to the right. There are expression methods such as visualization by visualization, visualization in the form of a bar graph, or drawing in the form of a pie chart, but are not limited thereto.

In claim 1 or 4 , the context of the user activity history recorded at the same time is obtained by reflecting the context of the user's operation content and the request content in the value of the classification ID and using this as a sort key. Can be accurately reproduced.

In claim 2 or 5 , when visualizing the ratio value, by visualizing a rectangle having a horizontal width corresponding to the ratio value corresponding to each classification ID in a color set for each classification ID, There are methods such as visualization with a pie chart and visualization with a bar chart, but are not limited thereto.

In claim 7 , a program capable of executing the inventions of claims 1 to 6 can be realized.

  According to the present invention, after adding a classification ID based on an acquisition place or function to a user activity history including time information recorded on various service providing servers or information terminals, each user's chronological order is added. By visualizing the integration of all activity histories, it is possible to understand in a simple manner what kind of work the user attaches importance to in a series of meaningful activities and in what order the services and functions are used properly. it can.

  In addition, by calculating the similarity between users using the feature vector calculated based on the usage time for each classification ID, the activity of a specific task such as online shopping or learning service is obtained. Users with similar trends can be extracted, and can be used for grasping activity trends related to success / failure and grasping user groups with different activity progresses.

  Therefore, it is possible to utilize grouping from a viewpoint different from the content-dependent one such as a hobby preference that has been conventionally performed.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a schematic configuration diagram showing a user activity history visualization / analysis system according to an embodiment of the present invention. In FIG. 1, the user 10 operates the information terminals 21-1 to 21-n in the information terminal group 20 to access the service providing servers 31-1 to 31-n in the service providing server group 30. Each operation such as browsing, bookmark operation, report publishing operation, report editing operation and communication operation is performed.

  At this time, the activity history of the user 10 in the information terminals 21-1 to 21-n is stored in the activity history DB 22-1 to 22-n, and the activity history of the service providing servers 31-1 to 31-n is stored in the activity history DB 32-1. 32-n respectively.

  When the user 10 or a third party requests the user activity history visualization analyzer 40 to display the activity status of the user 10, predetermined activity history DBs 22-1 to 22-n, 32-1 to record the activity history of the user 10. The corresponding activity history is individually read into the activity history aggregation unit 41 of the user activity history visualization analyzer 40 from 32-n.

  Based on the classification list 44, an ID is added to the read activity history for each history.

  The activity history aggregating unit 41 cuts out only the time information and the category ID from each activity history to which the category ID is added, and then integrates and sorts in time series. After the sorting is completed, the time difference between each activity history is further calculated. Through the above processing, the activity histories aggregated from different DBs are integrated based on the time information and the prescribed classification, and a time-series activity transition situation list for each user is generated.

  The generated time-series activity transition status list is imaged in the activity history visualization unit 43.

  In the activity history analysis unit 42, the time-series activity transition situation list generated for each user is re-sorted for each category ID, and the sum of elapsed time for each category ID is calculated. By dividing the sum of the elapsed times for each category ID obtained by the total activity time, the ratio of the time allocated to the activity corresponding to each category ID is calculated with respect to the total activity time. It is also possible to extract a user group having a similar feature vector by considering the obtained ratio value as a feature vector by similarity or cluster analysis.

  The activity history visualization unit 43 can also image the ratio of the time allocated to the activity corresponding to each classification ID.

  Next, the processing flow of this embodiment will be described with reference to FIG.

  First, the service usage will be described. First, the user 10 makes a service use request using the information terminal 21 (S1).

  The information terminal 21 records this service use request as an activity history record (S2).

  Further, the information terminal 21 accesses the service providing server 31 and makes a service use request (S3).

  The service providing server 31 records this service use request as an activity history record (S4).

  The service providing server 31 provides a service in response to a request from the information terminal 21 (S5), and the information terminal 21 provides a service in response to a request from the user 10 (S6).

  Next, the visualization time will be described. First, the user 10 or the like makes an activity history visualization request to the activity history visualization unit 43 (S7).

  The activity history visualization unit 43 makes a target activity history aggregation request to the activity history aggregation unit 41 (S8).

  The activity history aggregation unit 41 makes a target activity history collection request to the service providing server 31 and the information terminal 21 (S9).

  In response to the target activity history collection request, the service providing server 31 and the information terminal 21 transmit the target activity history to the activity history aggregating unit 41 by returning the target activity history (S10).

  The activity history aggregating unit 41 adds a classification ID to the target activity history transmitted from the service providing server 31 and the information terminal 21 (S11), integrates all activity history (S12), and sorts all activity history (S13). In addition, the time history activity transition status list is returned to the activity history visualization unit 43 (S14).

  The activity history visualization unit 43 generates an activity history visualization image for the time-series activity transition status list (S15), and transmits this image to the user 10 or the like by returning the activity history visualization result (S16).

  Next, a processing flow when the activity history analysis unit 42 performs activity history analysis will be described with reference to FIG.

  S1 to S6 and S8 to S14 are the same as the processing flow in FIG. 2, and in S7, an activity trend analysis request is made instead of an activity history visualization request.

  The activity history visualization unit 43 transmits the time-series activity transition status list received from the activity history aggregation unit 41 by returning the time-series activity transition status list to the activity history analysis unit 42 by returning the time-series activity transition status list (S21).

  Based on the time-series activity transition status list, the activity history analysis unit 42 performs tabulation by activity history classification ID (S22), activity history feature vector calculation (S23), and group extraction based on feature vectors (S24). The activity history analysis result is returned to the user 10 etc. (S25). In addition, after generating the image of the group extraction result based on the feature vector by the activity history visualization unit 43, the image may be transmitted to the user 10 or the like.

  Embodiments of the present invention will be described below with reference to the drawings.

  Here, a learning service is assumed that uses an information providing server 121 that provides various information via various networks and a learning support server 123 that supports learning using the information providing server 121.

  Each user 100 who is a learner basically accesses the learning support server 123 via the information terminal 111 and starts learning based on the presented task. Thereafter, during learning, it is possible to access information held by each information providing server 121 at any time, and the learner 100 uses the plurality of learning support functions provided by the learning support server 123 to solve the problem. After searching, accumulating, and reconfiguring, it can be compiled into a specified report format.

  At this time, the activity history during learning is individually recorded in the information providing server 121 and the learning support server 123 in different formats except for the recording time.

  When an activity status display of a learner 100 is requested during learning or after completion of learning, the corresponding activity history is individually activated from the activity history DBs 112, 122, and 124 including the activity of the learner 100. It is read into the history aggregating unit 131.

  At the same time, an ID is added to the read activity history based on the classification list 134 for each history line as shown in FIG. At this time, it is possible to arbitrarily set which classification ID is assigned to each of the servers 121 and 123, the function within the server, the information terminal 111, and the function within the information terminal. It is also possible to set a common classification ID in terms of meaning.

  In addition, by reflecting the context of the operation content of the user 100 corresponding to the classification ID, when the user activity history is sorted to create the time-series activity transition status list, the classification ID is also added to the time information. By using the sort key, it is possible to accurately reproduce the context of the user activity history recorded as being performed at the same time.

  For example, when the classification ID is represented by a numerical value, the “Web page” is assigned to the classification ID to be allocated by browsing the Web page and clicking the table of contents in the table of contents function of the learning support system (calling the table of contents page). If it is set so that “the classification ID of the viewing of the page” <“the classification ID of the click of the table of contents”, even if those histories are recorded at exactly the same time (in many cases, in millisecond units) Since the time difference is not recorded), the correct order relation (in this case, the Web page is browsed only after the table of contents is clicked) can be reproduced from the size of the classification ID. However, at this time, the classification ID does not necessarily need to be a numerical value, and may be a letter such as an alphabetic character or a hiragana character as long as sorting can be performed according to a certain rule.

  Here, as shown in FIG. 5, among the activity histories acquired from the activity history DB 122 of the information providing server 121, ID: 1 is added to information related to information browsing, and ID: 2 is added to information related to information retrieval. Among the activity histories acquired from the activity history DB 124 of the learning support server 123, ID: 3 is associated with the bookmark function, ID: 4 is associated with the report editing function, and ID: 4 is associated with the report publishing function. 5 and ID: 6 are added to those related to the communication function with other learners.

  At this time, by adding the descending order of the classification ID to the sorting subkey at the time of sorting, if the activity history is recorded at the same time, the higher the classification ID is displayed first, and the consistency of the history is maintained. Be drunk.

  Subsequently, the activity history aggregating unit 131 extracts and integrates only the time information and the category ID from each activity history to which the category ID is added and sorts them in time series.

  Here, “integration” means “combine (merge) into one table”. That is, the user activity histories recorded in the different activity history DBs are converted into a state having a common field of “time information” + “classification ID” by the assignment of the classification ID by the aggregation unit 131. Can be merged into a table.

  After the sorting is completed, a time difference between each history line (elapsed time until moving to the next user activity in each user activity history) is calculated.

  The activity histories collected from different DBs by the above processing are integrated based on the time information and the prescribed classification, and a time-series activity transition situation list as shown in Table 1 below is generated.

  The generated time series activity transition status list is imaged in the activity history visualization unit 133.

  For example, as shown in FIG. 6, by drawing a rectangle based on the corresponding elapsed time and the color of the classification ID to the right of the horizontal axis with the first line in the time series activity transition status list as the origin, The student's activity transition status is imaged in the form of color transition to the right of the horizontal axis. In FIG. 6, the learner's activity is gradually changing from collecting information from the information providing server 121 to report editing work and interaction with other learners, especially during the information search. You can easily see that the report is being edited from time to time.

  FIG. 7 shows an example in which the activity history for the classification IDs: 1 to 6 is visualized based on the above method.

  As described above, it is possible to extract learners having a similar activity tendency after grasping the rough transition of the activities of the learner 100. That is, the time series activity transition status list generated for each learner 100 as described above is re-sorted by the category ID in the activity history analysis unit 132, and the sum of elapsed time for each category ID is calculated. Here, the sum of elapsed time corresponding to six classification IDs is obtained for each learner. Furthermore, by dividing the total elapsed time of each category ID obtained by the total activity time, the ratio of the time allocated to the activity corresponding to each category ID with respect to the total activity time as shown in FIG. Is calculated.

  The group of learners 100 having similar feature vectors by considering the ratio value thus obtained as a feature vector and using similarity (using inner product value, L2, distance (straight distance between two points), etc.), cluster analysis, etc. Can be extracted.

  As a result, not only the progress of learning can be grasped, but also feedback regarding the learning method to the learner 100 itself (awareness of meta-learning), instruction for each group of learners 100 with similar activity tendencies, and the like can be performed.

  In the above embodiment, the learning service has been described. However, this visualization / analysis method is not applied only to the learning service, but the activity history such as online shopping, workflow analysis, etc. depending on the way of defining the classification. It can be applied to various scenes that can be acquired.

  Also, the target activity history itself is not only recorded by the service providing server 120 as described in the previous embodiment, but also recorded by the information terminal 111, and the date and time are recorded and digitized. Daily reports can be used in exactly the same way.

  In the present invention, some or all of the processing functions of the system shown in FIGS. 1 and 4 are configured as a program and realized using a computer, or the processing procedures shown in FIGS. 2 and 3 are configured as a program. Can be executed by a computer. In addition, a computer-readable recording medium such as a flexible disk, MO, ROM, or memory card can be used to store a program for realizing the processing function of each unit by the computer or a program for causing the computer to execute the processing procedure. It can be recorded on a CD, a DVD, a removable disk, etc., stored, provided, and distributed via a communication network such as the Internet.

1 is a schematic configuration diagram showing a user activity history visualization / analysis system. The processing flow figure in a user activity history visualization and analysis system. The processing flow figure in a user activity history visualization and analysis system. 1 is a schematic configuration diagram showing a user activity history visualization / analysis system. Explanatory drawing which shows addition of classification ID to an activity history. An explanatory view showing visualization of activity history. Explanatory drawing which shows the visualization example of activity history. Explanatory drawing which shows the feature vector and visualization of an activity tendency.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... User 20 ... Information terminal group 21 ... Information terminal 22 ... Activity history DB
30 ... Service providing server group 31 ... Service providing server 32 ... Activity history DB
DESCRIPTION OF SYMBOLS 40 ... User activity history visualization analyzer 41 ... Activity history aggregation part 42 ... Activity history analysis part 43 ... Activity history visualization part 44 ... Classification list 100 ... User 110 ... Information terminal group 111 ... Information terminal 112 ... Activity history DB
120 ... Service providing server group 121 ... Information providing server 122 ... Activity history DB
123 ... Learning support server 124 ... Activity history DB
DESCRIPTION OF SYMBOLS 130 ... User activity history visualization analyzer 131 ... Activity history aggregation part 132 ... Activity history analysis part 133 ... Activity history visualization part 134 ... Classification list

Claims (7)

An information terminal on which a user performs a service request operation, a service providing server that receives a user request from the information terminal and provides a service, and a user activity history that visualizes a user activity history in the information terminal and the service providing server A user activity history visualization / analysis method in a service providing system having a visualization / analysis device,
The information terminal adding time information to the user's operation content and recording it as a user activity history;
The service providing server adds time information to a user request content and records it as a user activity history;
The user activity history visualization / analysis device acquiring a user activity history recorded in the information terminal and the service providing server;
A step of attaching a classification ID to the acquired user activity history according to the operation content and the request content;
Sorting the acquired user activity history based on time information;
Calculating and adding elapsed time between user activity histories for the sorted user activity histories;
Creating a time-series activity transition status list for the operation content and the request content;
Visualizing the time series activity transition status list ,
The user activity history visualization / analysis method, wherein the user operation history and the request content are reflected in the classification ID value, and the classification ID value is also used as a key for the sorting. An information terminal on which a user performs a service request operation, a service providing server that receives a user request from the information terminal and provides a service, and a user activity history that visualizes a user activity history in the information terminal and the service providing server A user activity history visualization / analysis method in a service providing system having a visualization / analysis device,
  The information terminal adding time information to the user's operation content and recording it as a user activity history;
  The service providing server adds time information to a user request content and records it as a user activity history;
  The user activity history visualization / analysis device acquiring a user activity history recorded in the information terminal and the service providing server;
  A step of attaching a classification ID to the acquired user activity history according to the operation content and the request content;
  Sorting the acquired user activity history based on time information;
  Calculating and adding elapsed time between user activity histories for the sorted user activity histories;
  Creating a time-series activity transition status list for the operation content and the request content;
  Calculating a ratio value of activity time with respect to the total activity time for each classification ID from the created time-series activity transition status list;
  And visualizing the result of the ratio value calculation. A user activity history visualization / analysis method comprising: Obtaining a feature vector indicating an activity tendency of a user having the same dimension as the number of the classification IDs from a ratio value of the activity hours occupying with respect to all activity hours for each classification ID;
  The method for visualizing and analyzing a user activity history according to claim 2, further comprising: extracting users having similar activity tendencies using the feature vector calculated for each user. The user activity history recorded by the information terminal on which the user performs a service request operation with time information added to the user's operation content, and the service providing server that receives the user request from the information terminal and provides the service is the user A user activity history visualization / analysis device that receives user activity history recorded with time information added to the requested content and visualizes these user activity histories,
  Means for obtaining a user activity history recorded in the information terminal and the service providing server;
  Means for attaching a classification ID to the acquired user activity history according to the operation content and the request content;
  Means for sorting the acquired user activity history based on time information;
  Means for calculating and adding an elapsed time between user activity histories for the sorted user activity histories;
  Means for creating a time-series activity transition status list for the operation content and the request content;
  Means for visualizing the time-series activity transition status list,
  A user activity history visualizing / analyzing apparatus characterized in that the context of the user's operation content and request content is reflected in the value of the classification ID, and the value of the classification ID is also used as a key for the sorting. The user activity history recorded by the information terminal on which the user performs a service request operation with time information added to the user's operation content, and the service providing server that receives the user request from the information terminal and provides the service is the user A user activity history visualization / analysis device that receives user activity history recorded with time information added to the requested content and visualizes these user activity histories,
  Means for obtaining a user activity history recorded in the information terminal and the service providing server;
  Means for attaching a classification ID to the acquired user activity history according to the operation content and the request content;
  Means for sorting the acquired user activity history based on time information;
  Means for calculating and adding an elapsed time between user activity histories for the sorted user activity histories;
  Means for creating a time-series activity transition status list for the operation content and the request content;
  Means for calculating, from the created time-series activity transition status list, the ratio value of the activity time to the total activity time for each classification ID;
  Means for visualizing the result of the ratio value calculation, and a user activity history visualization / analysis device. Means for obtaining a feature vector indicating the activity tendency of the user in the same dimension as the number of the classification IDs from the ratio value of the activity time with respect to the total activity hours for each classification ID;
  6. The user activity history visualization / analysis apparatus according to claim 5, further comprising means for extracting users having similar activity tendencies using the feature vector calculated for each user. A user activity history visualization / analysis method or a user activity history visualization / analysis device according to any one of claims 1 to 6 described in a computer program and made executable .