JP2012173789A - Action classification apparatus, action classification method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To determine linkage of plural actions carried out at one time, and classify the synthetic action.SOLUTION: An information acquisition part 21 acquires action information including communication partner information and action object information of the action of an object person every scheduled time. A project information storage part 23 stores identification information of a project and project information including correspondence between activities object information to identify an object related to the project and member information. An action determination rule storage part 24 stores an action determination rule to establish correspondence between a condition of communication partner information and information targeted for an action and action kind to classify the action information. An action determination part 3 classifies action information in any of the action kinds according to the action determination rule by using the project information.

Description

  The present invention relates to a behavior classification device, a behavior classification method, and a program.

  2. Description of the Related Art There is known a technique for detecting position information of a person who performs a job or a job situation in a office or the like by using a camera image and automatically classifying job contents based on the detection result.

  For example, Patent Literature 1 describes an activity status overview system that grasps the activity status of a plurality of users regarding a certain project, including activities personally performed by a user. In the activity status bird's-eye view system of Patent Document 1, the contents of operations performed on the data for collaborative work and the data for personal work are managed as activity information, and a series of activities performed on a specific activity based on the activity information. Extract the operation details.

  Also, the work status information sharing system of Patent Document 2 determines a worker's task based on the worker's work status information and registered task information, and status information regarding the worker's task based on the work status information and the task information. Is generated.

  Patent Document 4 describes a technique for determining the homogeneity of an organization by analyzing the identity of behavior in the organization. The tissue homogeneity analyzer disclosed in Patent Document 4 uses information on group members to be analyzed, information about grouping log information using information on organization members to be analyzed, information on projects in business, information on job types, etc. Extract the homogeneity characteristics of the tissue from the log to be analyzed according to the conditions.

  In addition, Patent Literature 3 describes a user behavior induction system that classifies each behavior similar to the user behavior by, for example, clustering position vectors indicating the user behavior from behavior data detected by a sensor. . In Patent Document 5, a motion vector is obtained every predetermined time width (for example, 1 second), a similarity between the motion vector and each standard motion vector is obtained, and the similarity with the standard motion vector is calculated. It describes that an observation status vector as an element is created. In Patent Document 6, an action item classified by an observer who has observed a VTR image of a rat's behavior is compared with an image parameter extracted by image processing of the VTR image, and an algorithm for determining a rat's behavior is created. A method for automating behavior observation is described.

JP 2000-112893 A JP 2006-065436 A JP 2006-088251 A JP 2007-079666 A JP 2007-188287 A JP 11-296651 A

  In the techniques of Patent Documents 1, 2, or 4, if a plurality of actions (here, a meeting and a PC operation) are performed at the same time, such as a computer (PC) operation during a meeting, is the action independent? It is not possible to categorize what the action is as a whole. In Patent Documents 3, 5 and 6, it is only possible to classify the external motion of the observation target, and it is not possible to determine the relevance of a plurality of actions performed in parallel.

  The present invention has been made in view of the circumstances as described above, and even when a plurality of actions are performed in parallel, the relevance is determined, and the action classification for comprehensively classifying what action is performed An object is to provide a device, a behavior classification method, and a program.

The behavior classification device according to the first aspect of the present invention provides:
Action acquisition means for acquiring action information including communication partner information and action target information of the action of the target person at predetermined time intervals;
Project information storage means for storing project information including activity target information and member information for identifying objects related to the project in association with project identification information;
Rule storage means for storing an action determination rule for defining a correspondence relationship between the condition of the communication partner information and the action target information and an action type for classifying the action information;
In accordance with the action determination rule using the project information, action determination means for classifying the action information into any of the action types;
It is characterized by providing.

The behavior classification method according to the second aspect of the present invention includes:
A behavior classification method performed by a behavior classification device that classifies behavior information of a target person into a predetermined behavior type,
An action acquisition step for acquiring action information including communication partner information and action target information of the action of the subject at predetermined time intervals;
Using the project information including activity target information and member information for identifying an object related to the project associated with the project identification information, the communication partner information and the condition of the action target information, and the action information An action determination step for classifying the action information into one of the action types according to an action determination rule that defines a correspondence relationship with the action type to be classified;
It is characterized by providing.

A program according to a third aspect of the present invention provides a computer,
Action acquisition means for acquiring action information including communication partner information and action target information of the action of the target person for each predetermined time;
Project information storage means for storing project information including activity target information and member information for specifying an object related to the project in association with the project identification information;
Rule storage means for storing an action determination rule for defining a correspondence relationship between the condition of the communication partner information and the action target information and an action type for classifying the action information, and according to the action determination rule using the project information, Action determination means for classifying action information into one of the action types;
It is made to function as.

  According to the present invention, even when the subject is performing a plurality of activities at the same time, the relevance can be determined and the activities can be classified comprehensively.

It is a block diagram which shows the structural example of the action classification system which concerns on Embodiment 1 of this invention. 6 is a diagram illustrating an example of action log data according to Embodiment 1. FIG. 6 is a diagram illustrating an example of behavior information in Embodiment 1. FIG. 6 is a diagram illustrating an example of project information in Embodiment 1. FIG. 6 is a diagram illustrating an example of behavior determination rule information in Embodiment 1. FIG. 6 is a diagram illustrating an example of behavior classification information in Embodiment 1. FIG. 6 is a flowchart illustrating an example of behavior classification operation according to the first embodiment. It is a block diagram which shows the structural example of the action classification apparatus which concerns on Embodiment 2 of this invention. 10 is a diagram illustrating an example of a data structure of expanded behavior information in Embodiment 2. FIG. FIG. 10 is a diagram illustrating an example of expansion behavior information in the second embodiment. FIG. 10 is a diagram illustrating an example of behavior resource information in the second embodiment. It is a figure which shows the example of the action information weight data in Embodiment 2. FIG. 12 is a flowchart illustrating an example of behavior information similarity calculation according to the second embodiment. It is a figure which shows the example of the action determination rule information which concerns on Embodiment 3 of this invention. FIG. 20 is a diagram illustrating an example of action resource information in the third embodiment. It is a figure which shows the example of the action classification information which concerns on Embodiment 3. FIG. It is a block diagram which shows an example of the hardware constitutions of the action classification apparatus which concerns on embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the drawings, the same or equivalent parts are denoted by the same reference numerals.

(Embodiment 1)
FIG. 1 is a block diagram showing a configuration example of an action classification system according to Embodiment 1 of the present invention. The behavior classification system 1 includes a behavior classification device 2, a behavior log device 4, a terminal 5, and a server 6. The behavior log device 4 or the server 6 may be a part of the behavior classification device 2 or an attached device.

  The behavior classification system 1 collects the behavior log data of the target person, and classifies the behavior information generated from the behavior log data every predetermined time into behavior types of a predetermined viewpoint. The target person is a person who performs an action for which the action classification system 1 is to classify action information. In the present invention, the action refers to the action of a person accompanied by an intellectual activity, and particularly deals with the action of a business activity.

  The action information is composed of communication information and action target information of actions performed by the target person every predetermined time. The predetermined time may be a fixed time or may be a time segmented at the time when some event occurs. For example, one piece of behavior information may be generated from a certain amount of behavior such as every 10 minutes, 30 minutes, or 1 hour. Alternatively, for example, the length may be divided by an indefinite time such as from the start to the end of the meeting or from the start to the end of editing a certain document.

  The communication information includes communication partner information and field information in which communication is performed. Communication includes communication between people using telecommunications means in addition to people talking face to face. Communication includes, for example, general conversations, meetings, conferences, and interviews, telephone calls, video conferences, chats via communication networks, and reading and writing of e-mails.

  The communication partner information is information for identifying the partner with whom the target person communicated. Generally, it is the name of a communication partner, a mail address, or a handle name in SNS. The communication information may include a mailing list name (address).

  The field information in which communication is performed may be identification information of a place where a meeting or a conference is held, or information indicating standing in a living room, refresh room, or passage. In communication (e-mail, chat, etc.) in an information space, fields are an e-mail account, SNS (social network system), communication means, and the like.

  The action target information includes information for identifying an action target. The object of action may be, for example, an article to be processed or assembled, an object of experiment or observation, or a document to be created or viewed. The action target information may include information on an access place and / or a tool used. When the action target is information (electronic), the target identification information is a file path, the access location is a login terminal, and the tool used is a use application.

  In the present embodiment, the behavior classification device 2 acquires the behavior information of the target person every predetermined time, and from the communication information and the behavior target information, whether the behavior is a personal work or a meeting, and the behavior is a project. Classify behavior information from the perspective of whether it is relevant or not relevant.

In the present embodiment, a project refers to an activity performed for a certain purpose that can identify a member related to a target of an action related to the project. That the behavior of the subject is related to the project means that either of the following (1) or (2) has been performed, and that the behavior of the subject is not relevant to the project: It means that none of (2) is performed.
(1) Actions on subjects related to the project.
(2) Communication with members involved in the project.

  In the present invention, personal work means an action in which there is no communication partner. In personal work, there are actions that are not related to the actions related to the project. The personal work related to the project is (3) an action with respect to an object related to the project, with no communication partner. The personal work not related to the project is (4) an action with no communication partner with no action on the object related to the project. In personal work that is not related to the project, actions may be performed on objects other than those related to the project.

In the present invention, a meeting means an action in which a communication partner exists. Meetings include meetings that are not related to projects. There are the following two cases related to the project.
(5) The communication partner is a member related to the project, and no action is taken on the object related to the project.
(6) When a project involving a communication partner and a project involving an action target coincide.
A meeting not related to a project is an action when the communication partner information of the meeting is not empty, and members related to the project are not included or few in the communication partner information.

  The data collection unit 41 of the behavior log device 4 in FIG. 1 collects information on the subject of the subject's behavior, communication partner, and field information. For example, the data collection unit 41 is connected to the position sensor 43, the imaging device 44, and the microphone 45. For example, the position sensor 43 can detect the RF tag held by the subject and know the location of the subject. An RF tag detector may be provided at the entrance of the conference room or meeting corner, and the RF tag of the subject may be detected to determine whether the subject is in the conference room or the meeting corner.

  The position of the subject can also be extracted from an image photographed by the imaging device 44 installed indoors. Alternatively, the position may be detected by a GPS (Global Positioning System) receiver held by the subject. The action log device 4 acquires the position information of the target person by any one of these means. The action log device 4 can extract the target of the action performed by the subject in real space from the image taken by the imaging device 44.

  For example, the data collection unit 41 detects that the subject is communicating from the voice collected by the microphone 45 installed in the living room or the conference room or the microphone 45 held by the subject. The communication partner information and field information can be collected together with the position information of the target person and the conversation partner.

  The action log device 4 collects from the server 6 or the terminal 5 information on a file created, edited, or viewed by the target person, a terminal that accesses the file, and information on the application used. The action log device 4 also collects information that the target person has read and written the electronic mail from the server 6 and acquires communication partner information from the transmission address and the reception address of the electronic mail. In this case, account information obtained by reading and writing an electronic mail is used as field information.

  When the target person performs voice communication or text communication via the server 6, the action log device 4 collects partner information of the communication from the server 6. If the server 6 is a gateway to the Internet, information such as a URL (Uniform Resource Locator) of a WEB accessed by the target person is collected from the server 6.

  The terminal 5 is an information terminal operated by the subject person. In FIG. 1, a single information terminal that can be operated by a target person is shown. There may be two or more terminals 5 for one target person, or there may be a terminal 5 shared by a plurality of target persons. The terminal 5 is connected to a wired or wireless LAN. The terminal 5 may be placed at a certain place, or may be held and moved by the subject. The target person accesses the server 6 from the terminal 5 via the network 7.

  The terminal 5 includes an input unit 51, a display unit 52, and a communication unit 53. The input unit 51 includes an input device such as a keyboard and a mouse. Through the input unit 51, the subject inputs operations such as file creation / editing or browsing, and reading / writing of e-mails. The display unit 52 displays files and emails according to the operation of the subject. The terminal 5 is connected to the network 7 by the communication unit 53, and performs transmission / reception of a file to be created / edited or viewed and transmission / reception of an electronic mail. When performing voice communication from the terminal 5 via the network 7, the communication unit 53 transmits and receives voice data.

  The server 6 stores document files, and transmits / receives e-mails to / from other e-mail servers via the Internet or an intranet. The server 6 may also serve as a gateway to the Internet. In FIG. 1, a single server 6 is displayed so that a file server that collectively manages document files can be used as an e-mail server and a gateway. A server that provides a file server, an e-mail server, a gateway, voice communication, or character communication may be provided separately. Each server is not limited to one, and may be composed of two or more.

  The server 6 includes a communication unit 61, a processing unit 62, and a storage unit 63. The communication unit 61 includes a communication interface and the like, and communicates with the terminal 5 and the action log device 4 via the network 7.

  The processing unit 62 performs processing such as authentication of a target person who accesses the server 6 via the terminal 5, storage processing of a file to be created / edited or viewed, permission of search and transmission / reception, permission and provision of use of an application program, and the like. . Further, when the server 6 also serves as an electronic mail server or a gateway, the processing unit 62 performs these processes.

  The storage unit 63 stores application programs and data created or edited by applications. In the storage unit 63, a directory is formed corresponding to each project, and data used in the corresponding project is stored in these directories. In the case of an e-mail server, the storage unit 63 stores e-mail account information, e-mail data to be transmitted / received, time at which the e-mail is read / written, and history information to be transmitted / received. In the case of a gateway server, information such as the URL and time of access to the Internet is stored in correspondence with the target person's account.

  The behavior log device 4 collects behavior communication information and behavior target information in real space for each target person from data detected by the position sensor 43 and the imaging device 44. Further, from the server 6, action communication information and action target information in the information space for each target person are collected. The behavior log device 4 stores the collected communication information and behavior target information as behavior log data in the behavior log data storage unit 42.

  FIG. 2 shows an example of action log data in the first embodiment. FIG. 2 shows that, for example, a subject ID, a start date / time, an end date / time, a communication partner, a field, an access environment, a tool, and an action target are stored as one record as an action log. “Null” in the communication partner, access environment, tool, and action target indicates that there is no such information.

  The communication partner indicates the partner who is talking or the sender or destination of an e-mail to be read and written. The field is a place where the target person exists or an e-mail account. The access environment is the terminal 5 to be used, and the tool is a utilization application. In the case of the real space, the action target is, for example, an article or a material. In the case of the information space, for example, the action target is a directory (access file path) that stores a file to be accessed.

  FIG. 2 shows a part of the action log data of Y1 year M1 month D1. In FIG. 2, for example, it is shown that the subject Pa has acted on the action targets Da and Dc using the access environment Ta and the tool Sa in the field Fa at the time of h1: m1 to h1: m3. In addition, during the overlapping times h1: m2 to h1: m3, the subject Pa indicates that he / she has acted on the action target Dd using the tool Sb.

  The action classification device 2 in FIG. 1 classifies action information generated from action log data for each predetermined time into action types from a predetermined viewpoint. The behavior classification device 2 includes an information acquisition unit 21, a behavior information storage unit 22, a project information storage unit 23, a behavior determination rule storage unit 24, and a behavior determination unit 3.

  The information acquisition unit 21 receives the behavior log data stored in the behavior log data storage unit 42 of the behavior log device 4, generates behavior information for each predetermined time from the behavior log data, and the behavior information storage unit 22. To remember. For example, when generating action information at a certain time, the action log data extending over two or more fixed times is divided into each fixed time, and the action log data included in the fixed time is totaled. Generate behavior information.

  In addition, when a predetermined time is defined as a time segmented at the time of occurrence of some event, the time when the event of interest occurred is extracted from the behavior log data, and the behavior log data is aggregated for each time segmented by that time To do. For example, priority can be given to the start and end of a meeting and the start and end of file editing, and the time divided by the time of occurrence can be set as a predetermined time. Dividing action log data over two or more predetermined times is the same as in the case of generating action information at regular time intervals. The action log device 4 may generate the action information for each predetermined time from the action log data.

  FIG. 3 shows an example of behavior information in the first embodiment. The action information includes, for example, action ID, target person ID, communication partner, field, access environment, tool, and action target information. The action ID is a code that identifies a set of a target time for generating action information and a target person. However, two or more action log data may be merged into one action information, or one action log data may be divided into two action information.

  The communication partner or action target in FIG. 3 is the same as the action log. Communication information is composed of the communication partner and the field. The access environment, tool, and action target are included in the action target information.

  The communication information is preferably represented by a set of communication partner information and field information. When an e-mail is exchanged with another party Y while a meeting is being held with the other party X in the conference room, the communication information includes the combination {meeting party X, conference room} and {e-mail partner Y , Email account}. These have more information than the communication partner and field. For example, {meeting partner Y, conference room} and {email partner X, e-mail account}, where the other party is switched, are different from the previous communication information.

  In FIG. 3, the action information of action ID = A1 corresponds to the action log data of h1: m1 to h1: m3 of the subject Pa in FIG. The action information of action ID = A5 is a combination of action log data of h1: m1 to h1: m4 and action log data of h1: m2 to h1: m3 of the subject person Pc in FIG. .

  The project information storage unit 23 in FIG. 1 stores activity target information indicating an action target related to each project and project information including member information related to the project in association with the project. FIG. 4 shows an example of project information in the first embodiment. The project information includes a project name that identifies the project, a member name, and an activity target. FIG. 4 shows an example in which an activity target is defined by a file path in which files related to each project are stored.

  The behavior determination rule storage unit 24 stores a behavior determination rule that defines a correspondence relationship between a condition for determining whether or not behavior information is included in each behavior type and the behavior type for classifying the behavior information. Conditions for determining whether or not action information is included in each action type include conditions for communication partner information and action target information. The condition of the communication partner information includes conditions such as the number of elements of the communication partner information and the degree of coincidence between the communication partner information and the member information of each project. The condition of the action target information includes conditions such as the number of elements of the action target information and the degree of coincidence between the action target information and the activity target information of each project. The action determination rule includes a condition of a degree that the direction of the relationship between the communication partner information and the project information matches the direction of the relationship between the action target information and the project information.

  The degree of coincidence between the communication partner information and the member information of each project can be represented by, for example, a member relation vector arranged in the order of the project information with the degree of coincidence between the communication partner information and the member information for each project information as an element. The degree of coincidence between the action target information and the activity target information of each project can be represented by, for example, a related project vector in which the degree of coincidence between the action target information and the activity target information for each project information is arranged in the order of project information. . Expressed in this way, the degree of matching between the direction of the relationship between the communication partner information and the project information and the direction of the relationship between the action target information and the project information is expressed by the inner product of the member related vector and the related project vector. be able to.

Here, the degree of coincidence of each is
(The number of elements in the intersection of set X and set Y) / (number of elements in set Y)
(It is possible to swap set X and set Y) or
(The number of elements in the intersection of set X and set Y) / (number of elements in the union of set X and set Y)
Can be expressed as For example, the degree of coincidence between the communication partner information and the member information for each project information can be expressed by a value obtained by dividing the number of elements of the product of the communication partner information and the member information by the number of elements of the member information. Also, the degree of coincidence between the action target information and the activity target information for each project information is expressed by a value obtained by dividing the number of elements of the intersection of the action target information and the activity target information for each project information by the number of elements of the activity target information. be able to.

  If the communication partner information includes an e-mail address, voice communication destination or SNS handle name, etc., the information is converted into the same identification information stored in the project member information. The degree of coincidence may be calculated. Further, when the communication partner information includes a mailing list, the communication partner information is decomposed into addresses belonging to the mailing list, and further converted into the same identification information as the member information, and then the degree of coincidence is calculated. Alternatively, when the mailing list is included in the member information, the degree of coincidence may be calculated by adding the number of members included in the member list to the number of elements.

  FIG. 5 shows an example of action determination rule information in the first embodiment. In the behavior determination rule of FIG. 5, in addition to the condition and behavior type for determining whether or not behavior information is included in each behavior type, “related project” is defined corresponding to the behavior type. The “related project” is a standard for generating related project information representing a project related to the action specified by the action information classified into the action type.

  In the condition of FIG. 5, ComMems is a set in which elements of communication partner information of behavior information are arranged. ComPJ is a member related vector in which the degree of coincidence between communication partner information and member information for each project information is arranged in the order of project information. DirPJ is a related project vector in which the degree of coincidence between action target information and activity target information for each project information is arranged in the order of project information. The order of project information is arbitrary, but the order of ComPJ and DirPJ projects is the same. The absolute value (| DirPJ | etc.) represents the number of empty or non-zero elements in the set or vector. NULL indicates an empty set or nothing. Here, the norm of the vector is the number of non-zero elements of the vector. The norm of the vector may be represented by the square root of the sum of squares of vector elements (Euclidean distance from the start point to the end point of the vector).

  The action determination rule of FIG. 5 defines four action types: project-related personal work, project-unrelated personal work, project-related meeting, and project-unrelated meeting. The project-related personal work corresponds to a condition in which the related project vector is not a zero vector, and the inner product of the member-related vector and the related project vector is below a predetermined value α. The project-unrelated personal work corresponds to a condition in which the communication partner information is an empty set and the norm of the related project vector is 0 (below a predetermined value). The project-related meeting corresponds to a condition in which the member-related vector is not the 0 vector and the related project vector is the 0 vector, or the inner product of the member-related vector and the related project vector is a predetermined value α or more. The project unrelated meeting corresponds to a condition in which the communication partner information is not an empty set, and the norm of the member related vector and the norm of the related project vector are both 0 (not more than a predetermined value).

  The behavior determination unit 3 in FIG. 1 calculates the degree to which the behavior information is related to the project, and classifies the behavior information into a predetermined behavior type according to the behavior determination rule. The behavior determining unit 3 includes a member coincidence degree calculating unit 31, a target coincidence degree calculating unit 32, and an action type determining unit 33.

  The member coincidence calculation unit 31 calculates a member related vector from the communication partner information of the behavior information and the member information of the project information. The member related vector is calculated as follows, for example.

The target person of behavior information is Pros, the communication partner information is ComMems, the member information of the i-th project information is PMemsi, and the degree of coincidence between the communication partner information and the member information for each project,
If ComMems ≠ Null,
| {{Pros∪ComMems} ∩PMemsi} | / | {PMemsi} |,
0 if ComMems = NULL,
It is defined as In other words, in order to see the degree of coincidence between the members who are communicating and the members of the project, it is necessary to consider the target person, so the product set of the union of the target person and communication partner information and the member information of the project The ratio of the number of elements is considered as the degree of coincidence. ComPJ is a vector arranged in the order of projects with the degree of coincidence for each project information as an element.

For example, for the action ID = A2 in FIG. 3, the member coincidence degrees of the projects Ja, Jb, and Jc in FIG.
| {Pb, Pa, Pe} ∩ {Pa, Pb, Pc} | / | {Pa, Pb, Pc} | = 2/3,
| {Pb, Pa, Pe} ∩ {Pd, Pe, Pf} | / | {Pd, Pe, Pf} | = 1/3,
| {Pb, Pa, Pe} ∩ {Pa, Pb, Pe} | / | {Pa, Pb, Pe} | = 1
It becomes. The member-related vector ComPj is (2/3, 1/3, 1) and (2 / √14, 1 / √14, 3 / √14) when the length is normalized to 1.

  The target coincidence calculation unit 32 calculates a project related vector from the action target information of the action information and the activity target information of the project information. The degree of coincidence between the action information and the activity information for each project can be the ratio of the number of elements of the intersection of the action object information and the activity object information to the number of elements of the activity object information, like the member-related vector. . Here, for example, the following calculation is performed in consideration of the frequency of access.

The target coincidence calculation unit 32 calculates, as a project related vector (DirPJ), a vector arranged in the order of the project, with the number of accesses to a file under the directory (file path) for each project as an element. For example, the action targets Da, Dc, and Dd of action ID = A1 in FIG. 3 are “\\ shared server \ share \ project A \ report.doc”,
"\\ shared server \ share \ project A \ material.jpg",
“\\ shared server \ share \ project B \ report.doc”
If this is the case, the number of accesses to the files below the file path that is the activity target of the projects Pa, Pb, and Pc in FIG. The related project vector DirPJ is (2, 1, 0) and (2 / √5, 1 / √5, 0) when the length is normalized to 1.

  The action type determination unit 33 uses the member-related vector and the related project vector calculated from the action information to determine which action type condition of the action determination rule the action information applies to. Then, the action information is classified into action types of the conditions that apply.

For example, as described above, the action information of action ID = A2 in FIG.
ComPj = (2/3, 1/3, 1)
It is. Since the action target information is NULL, DirPJ = (0, 0, 0).
Since | ComPJ | ≠ 0 and | DirPJ | = 0, the condition of the project related meeting in FIG. Therefore, the action information of action ID = A2 is classified into action type = project related meeting. The related project is ComPJ.

  In the behavior determination rule of FIG. 5, the inner product ComPJ / DirPJ of the member related vector ComPJ and the related project vector DirPJ may be compared with a predetermined value α, so the lengths of the member related vector and the related project vector are normalized. It is desirable to make it. Or you may calculate with the cosine similarity of member related vector ComPJ and related project vector DirPJ.

  The behavior information determination unit stores information (behavior classification information) obtained by classifying behavior information into behavior types in the behavior information storage unit 22. FIG. 6 is a diagram illustrating an example of behavior classification information in the first embodiment. FIG. 6 shows that the action type and related project information are stored in association with the action ID. In the example of FIG. 6, the related project information is the project name of the largest element of DirPJ.

  FIG. 7 is a flowchart showing an example of behavior classification operation according to the first embodiment. Here, it is assumed that the behavior information has already been generated from the behavior log data and stored in the behavior information storage unit 22.

  The behavior determination unit 3 reads behavior information from the behavior information storage unit 22 (step S10). The member coincidence calculation unit refers to the project information, calculates the coincidence between the communication partner information and the member information for each project information, and calculates the member related vectors arranged in the order of the project information (step S11). The target coincidence calculation unit 32 calculates the degree of coincidence between the action target information and the activity target information for each project information, and calculates related project vectors arranged in the order of the project information using this as an element (step S12).

  The action type determination unit 33 refers to the action determination rule information and evaluates the condition of each action type using the action information, the calculated member-related vector, and the related project vector (step S13). Then, the action information is classified into action types that meet the conditions (step S14).

  The behavior determination unit 3 generates related project information corresponding to the behavior information in accordance with the criteria of the related project information corresponding to the classified behavior type (step S15). And action classification information is memorized by action information storage part 22 with related project information (Step S16).

  If there is action information that is not yet classified in the action information storage unit 22 (step S17; YES), the process returns to step S10 and repeats from the reading of the action information. If there is no action information that is not classified (step S17; NO), the action classification process is terminated.

  According to the behavior classification device 2 of the first embodiment, the behavior information is classified into one of the behavior types according to the behavior determination rule using the communication partner information and the behavior target information. Behavior information can be classified in terms of whether the behavior is related to the project or not. As a result, even if the target person performs a plurality of actions in parallel, the actions can be classified and understood.

  Further, in the first embodiment, it is possible to grasp the project most likely to be associated with the behavior information of the target person using the member related vector (ComPJ) and / or the related project vector (DirPJ).

(Embodiment 2)
FIG. 8 is a block diagram showing a configuration example of the behavior classification apparatus according to Embodiment 2 of the present invention. In the second embodiment, the similarity between the behavior information is calculated. The behavior classification device 2 according to the second embodiment includes a behavior information weight calculation unit 25, a behavior resource storage unit 26, and a behavior similarity calculation unit 27 in addition to the configuration of the first embodiment. Other configurations are the same as those of the first embodiment.

  In the second embodiment, related project information is used to calculate the similarity between behavior information. The processing up to the generation of the related project information is the same as in the first embodiment. Extended action information is defined by adding the corresponding project information corresponding to each action information. A description will be given from the case where related project information is generated for each action information in the configuration of the first embodiment.

  FIG. 9 is a diagram for explaining an example of the data structure of the expanded behavior information in the second embodiment. The expanded behavior information in FIG. 9 has a configuration in which a related project is added to the behavior information in FIG. Until the related project information is generated by the behavior information determination unit, NULL is set in the related project information.

  FIG. 10 shows an example of the enlarged action information in the second embodiment. FIG. 10 shows a state where the generated related project information is set. In the example of FIG. 10, the related project information that is not the 0 vector is a related project vector (DirPJ) or a member related vector (ComPJ). The behavior determination unit 3 in FIG. 8 stores the generated related project information in the behavior information storage unit 22 in association with the behavior information.

  The behavior classification device 2 according to the second embodiment calculates a predetermined similarity for each element of the expanded behavior information obtained by adding the corresponding related project information to the behavior information, and calculates the two behavior information from the weighted sum of the similarities for each element. The similarity is calculated. The weight for each element of the weighted sum can be set as appropriate. Embodiment 2 deals with the case where different weights are set for each action type.

  The behavior information weight calculation unit 25 calculates a similarity weight for each element of the expanded behavior information based on a predetermined criterion from the behavior resource information that defines an element of interest for each behavior type among the elements of the expanded behavior information. Typically, the weight of an element of interest for each action type is set larger than the weight of an element not included in the element of interest of that action type. For example, two large and small reference numerical values are set, the weight of an element of interest is set to a large numerical value, and the weight of an element not included in the element of interest is set to a small numerical value.

  The action resource storage unit 26 stores information for determining an element of interest of the expanded action information for each action type. FIG. 11 shows an example of action resource information in the second embodiment. In the action resource information of FIG. 11, elements to be focused on are determined for each action type of FIG.

  In the element of interest in FIG. 11, Pros represents a target person of action information, Apps represents a use program, FilePathes represents action target information (access file path), and Project represents related project information. ComMems is communication partner information. In FIG. 11, for example, the focus elements of action type = project-related personal work are Pros, Apps, FilePathes, and Project.

  The behavior information weight calculation unit 25 in FIG. 8 calculates a weight for each element using the behavior resource information. For example, let β be the weight for the element of interest of the expanded action information and γ be the weight for the other elements (β> γ).

  FIG. 12 is a diagram illustrating an example of action information weight data according to the second embodiment. Here, the weight for every element is given to each expansion action information. In this example, the weight of the element of interest is 2 and the weight of the other elements is 0.5.

  The behavioral similarity calculation unit 27 in FIG. 8 calculates a predetermined similarity for each element of the expanded behavior information, and calculates the similarity between the two behavior information from the weighted sum of the similarity for each element. The behavior similarity calculation unit 27 acquires expanded behavior information from the behavior information storage unit 22. Further, the weight for each element of each action type is acquired from the action information weight calculation unit 25. The behavior similarity calculation unit 27 calculates the similarity for each element of the expanded behavior information for the two expanded behavior information.

  The similarity for each element can be a value obtained by dividing the number of elements of the product set of both elements by the number of elements of the union of both elements. In this case, the denominator may be the number of one element, for example, the larger element number. In the case where the ratio of the number of matching elements is the similarity, in the case of only one element such as a target person, the value is 1 when matching and 0 when different.

  When the action target information is an access file path set, a common upper directory of elements can be calculated for each set, and the reciprocal of the directory distance between the common upper directories can be used as the similarity. For example, in the case of "\\ shared server \ share \ project A \ report.doc" and "\\ shared server \ share \ tmp \ tmp.txt", each one level up from the directory to which each accessed file belongs. Since the "\ share" directory and more are in common, the directory distance is 2 as the sum of the number of hierarchies. Taking the reciprocal of the directory distance, the similarity is 0.5.

  Further, when the related project information is a vector of the same dimension as in the case of the related project vector (DirPJ) or the member related vector (ComPJ), the cosine similarity can be set as the similarity. Alternatively, the inner product of the related project information can be set as the similarity.

  Regarding the similarity of communication information, the similarity of communication partner information and field information can be calculated independently. Since communication information may include information in real space and information space (e-mail, etc.), more accurately, the similarity should be calculated using a set of communication partner information and field information as a set element. Is desirable.

  The behavioral similarity calculation unit 27 calculates the similarity between the behavioral information by accumulating values obtained by multiplying the weights corresponding to the calculated similarities. In the case of the behavior information weight data in FIG. 12, when calculating the similarity of behavior information classified into different behavior types, the element having both large weights (= 2) has 4 as the square. When both weights are different, 2 × 0.5 = 1. In the case of an element having a small weight on both sides, 0.5 × 0.5 = 0.25.

Specifically, the case where the similarity between the action ID = A4 and the action ID = A5 in FIG. 10 is calculated will be described as an example. It is assumed that the action target Db is “\\ shared server \ share \ project A \ report.doc” and the action target Db is “\\ shared server \ share \ tmp \ tmp.txt”.
From the expanded behavior information of FIG. 10, the similarity of each element is the similarity of the subject = 1,
Communication partner information similarity = 0,
Field similarity = 0,
Access environment similarity = 1,
Tool similarity = 0
Similarity of action target (access file path set) = 0.5
Similarity of related project information = 0
It becomes. Multiplying each element by the action information weight of FIG. 12 and adding 1 · 2 · 2 + 0 · 0.5 · 0.5 + 0 · 0.5 · 0.5 + 1 · 0.5 · 0.5
+ 0 · 2 · 2 + 0.5 · 2 · 2 + 0 · 0.5 · 0.5 = 6.25
It becomes.

  FIG. 13 is a flowchart showing an example of the behavior information similarity calculation operation according to the second embodiment. Here, it is assumed that the behavior determination unit 3 has already classified behavior information into behavior types and generated related project information. In the behavior information similarity calculation in FIG. 13, the similarity between two designated behavior information is calculated.

  The behavior similarity calculation unit 27 reads the two specified enlarged behavior information from the behavior information storage unit 22 (step S20). The behavior information weight calculation unit 25 refers to the behavior resource information and assigns a weight for each element according to the behavior type of the two expanded behavior information (step S21).

  The behavior similarity calculation unit 27 calculates the similarity for each element of the two expanded behavior information (step S22). Then, the similarity between the two behavior information is calculated by multiplying the similarity for each element by the respective weights and adding them (step S23). The behavior similarity calculation unit 27 outputs the calculated similarity between the two expanded behavior information (step S24), and ends the process.

  According to the behavior classification device 2 of the second embodiment, it is possible to calculate the similarity of behavior information at different points in time for one target person and the similarity of the behavior information of different target persons. Furthermore, the weight of similarity calculation can be changed according to the action type into which the action information is classified. For this reason, it is possible to adjust the similarity for each action type. Further, it is possible to calculate the similarity between the behavior information classified into different behavior types, and to define the similarity different from the behavior information classified into the same behavior type.

  In the present embodiment, the weights β and γ are not limited to the above values, and can be set to arbitrary values. The weight is not limited to two stages but can be three or more stages. In addition, the element of interest of each action type is not limited to the above, and can be arbitrarily set.

  By using the similarity calculated by the behavior classification apparatus 2 of the present embodiment, for example, by defining the distance between the behavior information by the reciprocal of the similarity, a cluster analysis of the behavior information can be performed. For example, the result of cluster analysis can be used for business activity support in accordance with the behavioral characteristics of the target person or a group of the target person.

(Embodiment 3)
The behavior classification apparatus 2 according to the third embodiment classifies behavior information regarding a specific project (task) as an independent behavior type in addition to the four types of behaviors according to the first or second embodiment. The configuration of the behavior classification device 2 of the third embodiment is the same as the configuration of the second embodiment, but the behavior determination rule information is different.

  FIG. 14 is a diagram showing an example of behavior determination rule information according to Embodiment 3 of the present invention. In the behavior determination rule information in FIG. 14, the “system management” behavior type is added to the behavior type in FIG. 5. In the third embodiment, the behavior determination rule storage unit 24 in FIG. 8 stores the behavior determination rule information illustrated in FIG.

  Specific action types to be added to the four kinds of action types described in the first embodiment from the viewpoints of individual work or meeting and whether the action is related to the project or not can be arbitrarily set. . In the third embodiment, the field (Area) is a specific place (Fs) and there is an access environment (Host) (or the action target includes a specific activity target “Host”). , Action type "System management" has been added.

  In accordance with the addition of “system management” to the action type, the action type “system management” is also added to the action resource information. FIG. 15 is a diagram illustrating an example of action resource information according to the third embodiment. In this example, the target person (Pros), the access environment (Host), and the application (Apps) are set as the focus elements of the action type “system management”. In the third embodiment, the behavior resource storage unit 26 in FIG. 8 stores behavior resource information shown in FIG.

  In the third embodiment, the action type determination unit 33 in FIG. 8 refers to the action determination rule information shown in FIG. 14, and in addition to the above-described four types of action information conditions, the specific action type “system management”. It is determined whether or not the action information meets the condition. Here, there is a possibility that action information that applies to the condition of “system management” may also apply to other action types, for example, conditions of personal work not related to a project. In that case, it is possible to configure the behavior determination rule information so as to be prioritized and classified with respect to a specific behavior type. Or you may make it the structure which recognizes giving both of two action types to one action information.

  FIG. 16 shows an example of action classification information according to the third embodiment. In FIG. 16, in addition to the action classification information of FIG. 6, the action information of action ID = An is classified as “system management”, indicating that there is no related project information (NULL). This behavior information corresponds to the behavior information of behavior ID = An in FIG.

  As described above, according to the behavior classification apparatus 2 of the third embodiment, whether it is a personal work or a meeting, and whether the behavior is related to the project or not, the behavior types other than the four types of behavior Action information can be classified into specific action types. Also, for the behavior information of the specific behavior type, it is possible to adjust the weight for calculating the similarity to be used according to the feature of the behavior.

  FIG. 17 is a block diagram showing an example of a hardware configuration of the behavior classification apparatus according to the embodiment of the present invention. As shown in FIG. 17, the behavior classification device 2 includes a control unit 11, a main storage unit 12, an external storage unit 13, an operation unit 14, a display unit 15, an input / output unit 16, and a transmission / reception unit 17. The main storage unit 12, the external storage unit 13, the operation unit 14, the display unit 15, the input / output unit 16, and the transmission / reception unit 17 are all connected to the control unit 11 via the internal bus 18.

  The control unit 11 includes a CPU (Central Processing Unit) and the like, and executes processing for classifying behavior information according to a control program 19 stored in the external storage unit 13.

  The main storage unit 12 is composed of a RAM (Random-Access Memory) or the like, loads a control program 19 stored in the external storage unit 13, and is used as a work area of the control unit 11.

  The external storage unit 13 includes a non-volatile memory such as a flash memory, a hard disk, a DVD-RAM (Digital Versatile Disc Random Access Memory), a DVD-RW (Digital Versatile Disc ReWritable), and controls the processing of the behavior classification device 2. A program to be executed by the unit 11 is stored in advance, and data stored by the program is supplied to the control unit 11 in accordance with an instruction from the control unit 11, and the data supplied from the control unit 11 is stored.

  The operation unit 14 includes a pointing device such as a keyboard and a mouse, and an interface device that connects the keyboard and the pointing device to the internal bus 18.

  The display unit 15 includes a CRT (Cathode Ray Tube) or an LCD (Liquid Crystal Display), and displays a screen for outputting action classification information and related project information.

  The input / output unit 16 includes a serial interface or a parallel interface. When the behavior classification device 2 also serves as the behavior log device 4, a position sensor 43, an imaging device 44, a microphone 45, and the like are connected to the input / output unit 16.

  The transmission / reception unit 17 includes a network termination device or a wireless communication device connected to the network 7, and a serial interface or a LAN (Local Area Network) interface connected thereto. The transmission / reception unit 17 is connected to the terminal 5 and the server 6 via the network 7.

  The information acquisition unit 21, the behavior information storage unit 22, the project information storage unit 23, the behavior determination rule storage unit 24, the behavior determination unit 3, the behavior information weight calculation unit 25, the behavior resource of the behavior classification device 2 illustrated in FIGS. The processing of the storage unit 26 and the action similarity calculation unit 27 is performed by the control program 19 using the control unit 11, the main storage unit 12, the external storage unit 13, the operation unit 14, the display unit 15, the input / output unit 16, the transmission / reception unit 17, and the like. It is executed by processing it using as a resource.

  In addition, the above-described hardware configuration and flowchart are examples, and can be arbitrarily changed and modified.

  The central part that performs the action classification process composed of the control unit 11, the main storage unit 12, the external storage unit 13, the operation unit 14, the display unit 15, the input / output unit 16, the transmission / reception unit 17, the internal bus 18, etc. It can be realized by using a normal computer system regardless of a dedicated system. For example, a computer program for executing the above operation is stored and distributed in a computer-readable recording medium (flexible disk, CD-ROM, DVD-ROM, etc.), and the computer program is installed in the computer. Thus, the behavior classification device 2 that executes the above-described processing may be configured. Alternatively, the behavior classification device 2 may be configured by storing the computer program in a storage device of a server on a communication network such as the Internet and downloading the computer program from a normal computer system.

  Further, when the function of the behavior classification device 2 is realized by sharing of an OS (operating system) and an application program, or by cooperation between the OS and the application program, only the application program portion is stored in a recording medium or a storage device. May be.

  It is also possible to superimpose a computer program on a carrier wave and distribute it via a communication network. For example, the computer program may be posted on a bulletin board (BBS, Bulletin Board System) on a communication network, and the computer program distributed via the network. The computer program may be started and executed in the same manner as other application programs under the control of the OS, so that the above-described processing may be executed.

  A part or all of the above-described embodiment can be described as in the following supplementary notes, but is not limited thereto.

(Appendix 1)
Action acquisition means for acquiring action information including communication partner information and action target information of the action of the target person at predetermined time intervals;
Project information storage means for storing project information including activity target information and member information for identifying objects related to the project in association with project identification information;
Rule storage means for storing an action determination rule for defining a correspondence relationship between the condition of the communication partner information and the action target information and an action type for classifying the action information;
In accordance with the action determination rule using the project information, action determination means for classifying the action information into any of the action types;
A behavior classification apparatus comprising:

(Appendix 2)
The behavior determination rule includes the communication partner information, a member related vector arranged in the order of the project information with a degree of coincidence between the communication partner information and the member information for each project information, and the behavior target information and the The behavior classification apparatus according to appendix 1, including information on conditions related to related project vectors arranged in the order of the project information using a degree of coincidence with activity target information for each project information as an element.

(Appendix 3)
The action type is
Project related personal work corresponding to a condition, wherein the related project vector is not a zero vector, and the inner product of the member related vector and the related project vector is below a predetermined value;
Project-related personal work corresponding to a condition, wherein the communication partner information is an empty set, and the norm of the related project vector is equal to or less than a predetermined value,
The member related vector is not a 0 vector and the related project vector is 0, or an inner product of the member related vector and the related project vector is a predetermined value or more, and a project related meeting corresponding to a condition, and
The communication partner information is not an empty set, and the norm of the member related vector and the norm of the related project vector are both equal to or less than a predetermined value, the project unrelated meeting corresponding to the condition,
The action classification device according to appendix 2, characterized in that it includes the following four types.

(Appendix 4)
The behavior information includes field information with which the target person communicates, a login terminal name that accesses the behavior target, an application name used, and access file path information,
The project information includes a project name, the member information, and file path information to a shared directory of the project.
The action classification device according to any one of appendices 1 to 3, characterized in that:

(Appendix 5)
The behavior classification apparatus according to appendix 4, wherein the behavior type includes a type of a specific project behavior corresponding to a condition including the field information, the login terminal name or the application application name, or a combination thereof. .

(Appendix 6)
The behavior determination rule includes information on a reference for generating related project information representing a project related to the behavior specified by the behavior information classified into the behavior type,
The behavior determination means generates related project information representing a project related to the behavior specified by the behavior information according to a standard for generating the related project information of the behavior type into which the behavior information is classified.
The action classification device according to any one of appendices 1 to 5, characterized in that:

(Appendix 7)
For the two pieces of behavior information, a predetermined similarity for each element of the expanded behavior information obtained by adding the related project information corresponding to each behavior information to the behavior information is calculated, and the weighted sum of the similarities for each element is calculated. The behavior classification apparatus according to appendix 6, further comprising a similarity calculation unit that calculates the similarity between the two behavior information.

(Appendix 8)
A weight calculation unit that calculates a weight of similarity for each element of the expanded action information based on a predetermined criterion from action resource information that defines an element of interest of the expanded action information for each action type,
The similarity calculation unit calculates the weighted sum by multiplying the similarity for each of the two expanded action information by the weight for each element of both expanded action information calculated by the weight calculation unit,
The action classification device according to appendix 7, characterized in that:

(Appendix 9)
The weight calculating unit sets, for each of the action types, a weight of an element of the expanded action information to be focused set in the action resource information larger than a weight of an element not included in the element to be focused of the action type. The action classification device according to appendix 8, which is characterized.

(Appendix 10)
The action type is
Project related personal work corresponding to a condition, wherein the related project vector is not a zero vector, and the inner product of the member related vector and the related project vector is below a predetermined value;
Project-related personal work corresponding to a condition, wherein the communication partner information is an empty set, and the norm of the related project vector is equal to or less than a predetermined value,
The member related vector is not a 0 vector and the related project vector is 0, or an inner product of the member related vector and the related project vector is a predetermined value or more, and a project related meeting corresponding to a condition, and
The communication partner information is not an empty set, and the norm of the member related vector and the norm of the related project vector are both equal to or less than a predetermined value, the project unrelated meeting corresponding to the condition,
4 types of
The behavior resource information is
The elements of interest corresponding to the project-related personal work are the use application name, the access file path information, and the related project information,
The elements of interest corresponding to the project unrelated personal work are the use application name and the access file path information,
The element of interest corresponding to the project-related meeting is the communication partner information and the related project information,
The element of interest corresponding to the project unrelated meeting is the communication partner information.
The action classification device according to appendix 8 or 9, characterized in that.

(Appendix 11)
The action type includes the type of specific project action corresponding to a condition including the field information, the login terminal name or the application name used, or a combination thereof,
The behavior resource information includes the login terminal name and the application application name as an element of interest corresponding to the type of the specific project behavior,
The action classification apparatus according to Supplementary Note 10, wherein:

(Appendix 12)
A behavior classification method performed by a behavior classification device that classifies behavior information of a target person into a predetermined behavior type,
An action acquisition step for acquiring action information including communication partner information and action target information of the action of the subject at predetermined time intervals;
Using the project information including activity target information and member information for identifying an object related to the project associated with the project identification information, the communication partner information and the condition of the action target information, and the action information An action determination step for classifying the action information into one of the action types according to an action determination rule that defines a correspondence relationship with the action type to be classified;
A behavior classification method comprising:

(Appendix 13)
The behavior determination rule includes the communication partner information, a member related vector arranged in the order of the project information with a degree of coincidence between the communication partner information and the member information for each project information, and the behavior target information and the 13. The behavior classification method according to appendix 12, including information on conditions related to related project vectors arranged in the order of the project information using a degree of coincidence with activity target information for each project information as an element.

(Appendix 14)
The behavior determination step includes
Project related personal work corresponding to a condition, wherein the related project vector is not a zero vector, and the inner product of the member related vector and the related project vector is below a predetermined value;
Project-related personal work corresponding to a condition, wherein the communication partner information is an empty set, and the norm of the related project vector is equal to or less than a predetermined value,
The member related vector is not a 0 vector and the related project vector is 0, or an inner product of the member related vector and the related project vector is a predetermined value or more, and a project related meeting corresponding to a condition, and
The communication partner information is not an empty set, and the norm of the member related vector and the norm of the related project vector are both equal to or less than a predetermined value, the project unrelated meeting corresponding to the condition,
14. The behavior classification method according to appendix 13, wherein the behavior information is classified into any one of the following four types.

(Appendix 15)
The behavior information includes field information with which the target person communicates, a login terminal name that accesses the behavior target, an application name used, and access file path information,
The project information includes a project name, the member information, and file path information to a shared directory of the project.
15. The behavior classification method according to any one of appendices 12 to 14, characterized in that:

(Appendix 16)
The behavior classification method according to appendix 15, wherein the behavior type includes a type of a specific project behavior corresponding to a condition including the field information, the login terminal name or the application application name, or a combination thereof. .

(Appendix 17)
The behavior determination rule includes information on a reference for generating related project information representing a project related to the behavior specified by the behavior information classified into the behavior type,
The behavior determination step generates related project information representing a project related to the behavior specified by the behavior information according to a criterion for generating the related project information of the behavior type into which the behavior information is classified.
The behavior classification method according to any one of appendices 12 to 16, characterized in that:

(Appendix 18)
For the two pieces of behavior information, a predetermined similarity for each element of the expanded behavior information obtained by adding the related project information corresponding to each behavior information to the behavior information is calculated, and the weighted sum of the similarities for each element is calculated. The behavior classification method according to appendix 17, further comprising a similarity calculation step of calculating a similarity between the two behavior information.

(Appendix 19)
A weight calculating step of calculating a weight of similarity for each element of the expanded action information based on a predetermined reference from the action resource information that defines an element of the expanded action information to be focused on for each action type;
The similarity calculating step calculates the weighted sum by multiplying the similarity for each of the two expanded action information by the weight for each element of both expanded action information calculated by the weight calculation unit,
The behavior classification method according to supplementary note 18, characterized by:

(Appendix 20)
In the weight calculation step, for each of the action types, the weight of the element of the expanded action information to be noticed determined in the action resource information is set larger than the weight of the element not included in the element of the action kind to be noticed. The behavior classification method according to supplementary note 19, which is a feature.

(Appendix 21)
The action type is
Project related personal work corresponding to a condition, wherein the related project vector is not a zero vector, and the inner product of the member related vector and the related project vector is below a predetermined value;
Project-related personal work corresponding to a condition, wherein the communication partner information is an empty set, and the norm of the related project vector is equal to or less than a predetermined value,
The member related vector is not a 0 vector and the related project vector is 0, or an inner product of the member related vector and the related project vector is a predetermined value or more, and a project related meeting corresponding to a condition, and
The communication partner information is not an empty set, and the norm of the member related vector and the norm of the related project vector are both equal to or less than a predetermined value, the project unrelated meeting corresponding to the condition,
4 types of
The behavior resource information is
The elements of interest corresponding to the project-related personal work are the use application name, the access file path information, and the related project information,
The elements of interest corresponding to the project unrelated personal work are the application name used and the access file path information,
The element of interest corresponding to the project-related meeting is the communication partner information and the related project information,
The element of interest corresponding to the project unrelated meeting is the communication partner information.
The action classification method according to appendix 19 or 20, characterized in that:

(Appendix 22)
The action type includes the type of specific project action corresponding to a condition including the field information, the login terminal name or the application name used, or a combination thereof,
The behavior resource information includes the login terminal name and the application application name as an element of interest corresponding to the type of the specific project behavior,
The behavior classification method according to Supplementary Note 21, wherein

(Appendix 23)
Computer
Action acquisition means for acquiring action information including communication partner information and action target information of the action of the target person for each predetermined time;
Project information storage means for storing project information including activity target information and member information for specifying an object related to the project in association with the project identification information;
Rule storage means for storing an action determination rule for defining a correspondence relationship between the condition of the communication partner information and the action target information and an action type for classifying the action information, and according to the action determination rule using the project information, Action determination means for classifying action information into one of the action types;
A program characterized by functioning as

DESCRIPTION OF SYMBOLS 1 Action classification system 2 Action classification apparatus 3 Action determination part 4 Action log apparatus 5 Terminal 6 Server 7 Network 11 Control part 12 Main memory part 13 External storage part 14 Operation part 15 Display part 16 Input / output part 17 Transmission / reception part 18 Internal bus 19 Control program 21 Information acquisition unit 22 Behavior information storage unit 23 Project information storage unit 24 Behavior determination rule storage unit 25 Behavior information weight calculation unit 26 Behavior resource storage unit 27 Behavior similarity calculation unit 31 Member coincidence calculation unit 32 Target coincidence calculation Unit 33 Action type determination unit 41 Data collection unit 42 Action log data storage unit 43 Position sensor 44 Imaging device 45 Microphone 51 Input unit 52 Display unit 53 Communication unit 61 Communication unit 62 Processing unit 63 Storage unit

Claims (10)

Action acquisition means for acquiring action information including communication partner information and action target information of the action of the target person at predetermined time intervals;
Project information storage means for storing project information including activity target information and member information for identifying objects related to the project in association with project identification information;
Rule storage means for storing an action determination rule for defining a correspondence relationship between the condition of the communication partner information and the action target information and an action type for classifying the action information;
In accordance with the action determination rule using the project information, action determination means for classifying the action information into any of the action types;
A behavior classification apparatus comprising:   The behavior determination rule includes the communication partner information, a member related vector arranged in the order of the project information with a degree of coincidence between the communication partner information and the member information for each project information, and the behavior target information and the The behavior classification apparatus according to claim 1, further comprising information on conditions related to related project vectors arranged in the order of the project information using a degree of coincidence with activity target information for each project information as an element. The action type is
Project related personal work corresponding to a condition, wherein the related project vector is not a zero vector, and the inner product of the member related vector and the related project vector is below a predetermined value;
Project-related personal work corresponding to a condition, wherein the communication partner information is an empty set, and the norm of the related project vector is equal to or less than a predetermined value,
The member related vector is not a 0 vector and the related project vector is 0, or an inner product of the member related vector and the related project vector is a predetermined value or more, and a project related meeting corresponding to a condition, and
The communication partner information is not an empty set, and the norm of the member related vector and the norm of the related project vector are both equal to or less than a predetermined value, the project unrelated meeting corresponding to the condition,
The behavior classification apparatus according to claim 2, comprising the following four types. The behavior information includes field information with which the target person communicates, a login terminal name that accesses the behavior target, an application name used, and access file path information,
The project information includes a project name, the member information, and file path information to a shared directory of the project.
The behavior classification device according to claim 1, wherein the behavior classification device is a device. The behavior determination rule includes information on a reference for generating related project information representing a project related to the behavior specified by the behavior information classified into the behavior type,
The behavior determination means generates related project information representing a project related to the behavior specified by the behavior information according to a standard for generating the related project information of the behavior type into which the behavior information is classified.
The behavior classification device according to claim 1, wherein the behavior classification device is a device.   For the two pieces of behavior information, a predetermined similarity for each element of the expanded behavior information obtained by adding the related project information corresponding to each behavior information to the behavior information is calculated, and the weighted sum of the similarities for each element is calculated. The behavior classification apparatus according to claim 5, further comprising a similarity calculation unit that calculates a similarity between the two behavior information. A weight calculation unit that calculates a weight of similarity for each element of the expanded action information based on a predetermined criterion from action resource information that defines an element of interest of the expanded action information for each action type,
The similarity calculation unit calculates the weighted sum by multiplying the similarity for each of the two expanded action information by the weight for each element of both expanded action information calculated by the weight calculation unit,
The behavior classification apparatus according to claim 6.   The weight calculating unit sets, for each of the action types, a weight of an element of the expanded action information to be focused set in the action resource information larger than a weight of an element not included in the element to be focused of the action type. The behavior classification device according to claim 7, wherein the behavior classification device is characterized in that: A behavior classification method performed by a behavior classification device that classifies behavior information of a target person into a predetermined behavior type,
An action acquisition step for acquiring action information including communication partner information and action target information of the action of the subject at predetermined time intervals;
Using the project information including activity target information and member information for identifying an object related to the project associated with the project identification information, the communication partner information and the condition of the action target information, and the action information An action determination step for classifying the action information into one of the action types according to an action determination rule that defines a correspondence relationship with the action type to be classified;
A behavior classification method comprising: Computer
Action acquisition means for acquiring action information including communication partner information and action target information of the action of the target person for each predetermined time;
Project information storage means for storing project information including activity target information and member information for specifying an object related to the project in association with the project identification information;
Rule storage means for storing an action determination rule for defining a correspondence relationship between the condition of the communication partner information and the action target information and an action type for classifying the action information, and according to the action determination rule using the project information, Action determination means for classifying action information into one of the action types;
A program characterized by functioning as

Images