JP2007072549A - Communication analysis device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform analysis of distribution of topics or trace of transition of a specific topic using the similarity of statement contents. <P>SOLUTION: A statement content analysis part 105 analyzes a statement and creates a user profile (graph) related to the content thereof. A task structure analysis part 105 generates a task list of each user based on user information registration tasks. A profile similarity determination part 108 calculates a similarity of profile (statement content) between users. A distance calculation part 109 calculates a task distance of the shortest route traced in reference to a task map. A similarity tabulation part 110 tabulates distribution of similarity (the number or average value of similar statement users), and an analysis result output part 111 outputs the result. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a technique for analyzing communication, and in particular, can analyze a distribution of similar communication based on a distance between members.

  With the development of electronic communication technology, it has become important to analyze communication. In particular, it is extremely important to understand the transition of topics and overall trends in an organization. For example, it is important to determine that an organization's activities are stiff and take some action. The present invention introduces the concept of distance between members of an organization and tries to analyze the distribution of communication based on the distance.

Patent documents 1 to 3 are related to the present invention. Patent Document 1 proposes a technique that enables access rights to be set and shared with personal network data. Patent Document 2 and Patent Document 2 propose a technique for distributing information that is not used by the user for communication and a hot topic using a pipe that expresses the connection between users. However, in the proposal of such patent document, since the connection data or pipe information is intended to express or match the connection between individuals, it is difficult to grasp the transition of topics in the organization and the overall trend. there were.
JP 2001-175680 A JP 2001-067359 A JP 2001-101139 A

  The present invention has been made in consideration of the above circumstances, and provides a communication analysis technique for analyzing the distribution of topics by using the degree of similarity of speech contents as an index in a communication system or for tracking changes in specific topics. The purpose is to do.

  The principle configuration example of the present invention is a communication system that determines the rigidity of an organization by analyzing the distribution of topics using the degree of similarity of speech content as an index or by tracking the transition of specific topics. It is. In other words, in a communication system, the change in the time distribution of the message distribution with the similarity as an index is analyzed to determine whether the topic has become rigid in the organization.

  In this configuration example, the structural distance between tasks is defined from the registration status in a virtual dialogue space (also called a task), and the similarity of the user's speech content is analyzed for each distance. Furthermore, by analyzing the temporal change in the similarity, the change in the spatial and temporal distribution of similar utterances is presented. It is also possible to limit the topic to be analyzed by specifying one or more keywords. In the following, “task” may be used in a limited sense as a group that can be defined by an electronic communication tool.

  The dialogue space is a mailing list, an electronic bulletin board, a web space (including a web log), or the like, but is not limited thereto. Not only an electronic dialogue space but also a physical dialogue space as long as it can be processed electronically.

  The distance between tasks in this configuration example will be described. In this configuration example, it is assumed that users registered in a plurality of tasks are in charge of each task. That is, it has a role of relating tasks. The task to which I belong is distance 1, the task that can be reached via one of them is distance 2,. . . It is defined as The distance between a predetermined user and another user is determined based on the registration relationship with the shortest reachable task. For example, in the example of FIG. 1, “user A / B” is registered in “task 1”, and “user B / C / D / E” is registered in “task 2”. “User E / F” is registered in “Task 3”. “User B” is registered in “task 1” and “task 2”, and is considered to be between “user A” and “user C / D / E” registered in each. Via "user B" and "user E", "user A" and "user F" are connected (the task 1 and the task 3 that are structurally irrelevant are related). The distance from “user A” to “task 1” (or other members belonging to “task 1”) is 1, and the distance to “task 2” (or members other than user B belonging to “task 2”) Is 2, and the distance to “task 3” (or a member other than user E belonging to “task 3”) is 3.

  In this way, distances between users, between user tasks, and between tasks are defined.

  FIG. 2 is a line graph for each user showing the relationship between the distance between tasks connected by the user and the number of accessible users (cumulative). In this example, all users are accessible at a distance of 6.

  In this configuration example, for each user or each of one or more selected users, the similarity (average value) of the contents of the remarks of the user located at each distance is displayed ((A) of FIGS. 3 to 5), For each distance, the number of similar utterance users among the users arranged at the distance is displayed ((B) of FIGS. 3 to 5).

  The example of FIG. 3 is a case where the similarity is determined for all extracted words without narrowing down the phrase, and the similarity tends to decrease as the distance increases. Further, although the distribution of the number of similar utterance users depends on the similarity of the reference threshold value, in the example of FIG. 3, most users are similar utterance users at all distances.

  The example of FIG. 4 is a case where the similarity is determined only for a special word (for example, “community”), and the similarity distribution has irregularities. The number of similar utterance users is reduced as a whole, and the number is also uneven. In the embodiment described later, the similarity is discriminated using an adjacent phrase of the phrase selected in the phrase graph, but the present invention is not limited to this.

  The example of FIG. 5 is a case where the similarity is determined by limiting to relatively frequently used phrases (for example, “document” “service”). In this example, the similarity is the same as the overall judgment in FIG. 3, and the distribution tendency of the number of similar utterance users is also similar. However, the number of users with respect to the absolute number is small.

  For example, an analysis as shown in FIG. 6 can be performed using the similarity of each task or the content of the user's utterance. In other words, (1) when the task of interest (or its user) is in the same task and the similarity of the content of the statement is high, the sound is sound. (2) When the target task (or its user) is in the same task and the similarity of the content of the remark is low, there is no dialogue (the remark is free to say). (3) If the task of interest (or its user) is a nearby task and the similarity of the content of the statement is high, there is a tendency to homogenize (for example, there is close cooperation). (4) If the task of interest (or its user) is a nearby task and the similarity of the content of the statement is low, there is a specialization tendency (for example, high expertise). (5) If the task of interest (or its user) is a task other than the same / neighbor and the degree of similarity of the content of the statement is high, new forces tend to emerge (for example, potential collaboration opportunities) There is). (6) If the task of interest (or its user) is a task other than the same / neighbor, and the degree of similarity of the contents of speech is low, it is average.

  The task (speech space) may be newly constructed on the premise of the communication analysis of the present invention, or may use an existing electronic bulletin board or mailing list thread or web access. In the case of electronic bulletin boards and mailing lists, the task structure is associated with the structure provided by each system. In the case of web access, a user profile may be created from the contents of the browsing page, and the task structure may be changed to a link structure.

  The present invention will be further described. In the following, for convenience of explanation, the corresponding portions of the embodiments are illustrated using symbols, but this is not intended to limit the technical scope of the invention.

  According to one aspect of the present invention, in order to achieve the above-mentioned object, the communication analysis apparatus: a similarity for determining a similarity between communication contents of members of an organization and communication contents of other members of the organization A determination unit (profile similarity determination unit 108); an inter-member distance calculation unit (distance calculation unit 109) for calculating a distance between a member of the organization and another member; and calculated by the inter-member distance calculation unit Totaling means (similarity totaling unit 110) for totalizing the similarity of the similarity determination means based on the distance is provided.

  The organization is typically a group such as a group such as a company or a department, a section, or the like, but is not limited thereto, and may be a project that spans an organization such as a company. Further, the organization is not limited to those that can be grasped as social realities, and may be, for example, a virtual or electronic community such as a mailing list or a bulletin board.

  In this configuration, it is possible to grasp the distribution of similar communications based on the distance between members.

  In this configuration, the similarity determination unit typically determines the similarity between the communication contents based on words included in the communication contents. In a specific example, words are extracted by morphological analysis of communication, a word network (graph) for each user is constructed, and the similarity between the word networks is calculated.

  Further, the inter-member distance calculation means calculates the inter-member distance based on, for example, the tissue part to which the member belongs. The organization part is, for example, a virtual space (task, a group that can be defined by an electronic communication tool). Tasks can be registered in a virtual space, and their members can communicate with each other using a mailing list, electronic bulletin board, Weblog, or the like. A member is registered in the task. For example, the distance from the member of interest to the other member is determined according to the number N of the smallest tissue portions on the route that connects the two members belonging to each tissue portion and reaches the other member from the member of interest. It is done. For members a, b, and c, when a and b belong to task A and b and c belong to task B, the distance is 2 because tasks A and B are interposed between a and c.

  Moreover, the said totaling means totals the number of the similarities in which a similarity degree exceeds a predetermined threshold value, for example, or totals statistical values, such as an average value.

  Moreover, it is preferable to provide a display means for displaying the result of counting by the counting means.

  Moreover, you may further provide the determination means which determines the rigidity of the said structure | tissue based on the total result of the said total means.

  Moreover, it is preferable to input information on a member serving as a reference for distance, information for specifying a time range of communication to be determined, and a keyword for narrowing down communication by an input unit.

  According to another aspect of the present invention, in order to achieve the above-described object, the communication analysis apparatus determines: a degree of similarity between communication contents of members of an organization and communication contents of other members of the organization A similarity determination unit that calculates a distance between a member of the organization and another member based on a group in which the member is registered, and a distance calculation unit that calculates the distance between the members. Aggregation means for aggregating the similarity of the similarity determination means on the basis of the distance is provided.

  The group may be any group as long as members are registered, but is typically a group that can be defined by an electronic communication tool.

  Even in this configuration, it is possible to grasp the distribution of similar communication based on the distance between members.

  According to still another aspect of the present invention, in order to achieve the above-mentioned object, the tissue stiffening analysis apparatus is provided: similarity between communication contents of members of an organization and communication contents of other members of the organization Similarity determining means for determining the degree; inter-member distance calculating means for calculating the distance between the members of the organization and other members; and the similarity based on the distance calculated by the inter-member distance calculating means Totaling means for totalizing the similarity of the determining means; display means for displaying the totaling result of the totaling means; and determining means for determining the rigidity of the tissue based on the totaling result of the totaling means.

  In this configuration, it is possible to grasp the distribution of similar communication on the basis of the distance between the members and recognize the rigidity of the organization.

  The present invention can be realized not only as an apparatus or a system but also as a method. Of course, a part of the invention can be configured as software. Of course, software products used to cause a computer to execute such software are also included in the technical scope of the present invention.

  These and other aspects of the invention are set forth in the appended claims and will be described in detail below with reference to examples.

  According to this invention, it is possible to analyze the distribution of topics using the degree of similarity of the content of an index as an index, or to track the transition of a specific topic, and more specifically, to determine whether the organization has become rigid. it can. In a specific example, the analyst can grasp the variation of the spatial and temporal distribution of similar statements, and it becomes easy to determine whether the topic is stiff in the organization. In a more specific example, the transition of a specific topic can be tracked by specifying the content of the utterance to be analyzed with one or more keywords, allowing analysis of the stiffening process focusing on a specific topic in the organization. It becomes.

  Examples of the present invention will be described below.

  FIG. 7 shows the communication analysis system 100 according to the embodiment of the present invention as a whole, and FIGS. 8 and 9 schematically show the preprocessing phase and the analysis phase of the communication analysis system 100.

  The communication analysis system 100 of this embodiment analyzes the communication on the communication system 200 in cooperation with the communication system 200. In the communication system 200, a dialogue space (task) is set and registered, and communication is performed between registered members. Communication may be web-based, electronic bulletin board, or mailing list. Any member can be used as long as members can be registered and communication is possible between the registered members. In addition to the member who makes a statement, there may be a person who can subscribe and view the content of the statement. However, such persons are not subject to communication analysis.

  The communication analysis system 100 may be mounted on one computer 300 together with the communication system 200, or may be mounted on another computer. In order to implement the communication analysis system 100 in the computer 300, the program 301 may be installed in the computer 300. The functional blocks of the communication analysis system 100 described below are realized by cooperating hardware resources and software resources of the computer 300.

  First, an overview of the overall processing flow will be described with reference to FIGS. 8 and 9.

  FIG. 8 shows an outline of the preprocessing phase of the communication analysis system 100. In this figure, the log 201 of the communication system 200 holds a message log and a task information log. In the preprocessing, a content map, a user map, and a task map are generated from the message log and the task information log. Then, a user profile is generated from the content map and the user map. The task map is a list of users belonging to each task. The user map is a list of tasks to which each user belongs. The user profile is a profile generated for each user from the posted content of the user. The content map is a list of phrases extracted from the message by, for example, morphological analysis. The user profile is, for example, a graph of words / phrases extracted from all messages of the user of interest, but is not limited thereto.

  FIG. 9 shows the outline of the analysis phase of the communication analysis system 100. In this figure, the analysis phase traces the user network (FIG. 1) and analyzes for the same distance, and the topic distribution from the past utterance contents of each person. (Similarity, similar utterance user) is calculated. When analyzing the content of a statement between one user (uid-001) and another user, first, the user profile of the user (uid-001) is extracted. A task (tid-001) to which the user belongs is extracted with reference to the user map of the user (uid-001), and another user (uid-002) belonging to the task is specified. This corresponds to the case where the task distance is 1, but when the task distance is 2 or more, the target user may be reached by tracing another task via another user. Next, the user profile of the other user (uid-002) is extracted, and the degree of similarity with the profile of the user of interest (uid-001) extracted earlier is calculated.

  The similarity for each distance is calculated, for example, by calculating the average. In addition, the number of users exceeding a predetermined threshold is tabulated for each distance. Aggregation can be performed in various other ways.

  Next, returning to FIG. 7, the configuration of the communication analysis system of the embodiment will be described.

  In FIG. 7, the communication analysis system 100 is linked to the communication system 200. The communication analysis system 100 may be a module of the communication system 200. The communication analysis system 100 includes a preprocessing unit 101, a user profile database 102, a task structure database 103, an analysis unit 104, and the like. The pre-processing unit 101 includes an utterance content analysis unit 105 and a task structure analysis unit 106. The analysis unit 104 includes an analysis condition input unit 107, a profile similarity determination unit 108, a distance calculation unit 109, a similarity calculation unit 110, an analysis result output unit 111, and the like.

  The preprocessing unit 101 performs the preprocessing phase described with reference to FIG. 8 as a whole. The speech content analysis unit 105 of the preprocessing unit 101 performs, for example, processing as shown in FIG. 11 as a whole. A word profile extracted from each message is constructed into a graph to construct a user profile, and the processing is as follows.

[Step S10]: A new user table is input to the user map (UTAB).
[Step S11]: The next message is substituted into M.
[Step S12]: If M is completed, the process ends. If not, the process proceeds to step S13.
[Step S13]: The caller is assigned to U. The user profile is searched with U, and the hit user profile is substituted into P _t−1 . Remarks time of the message is assigned to the t _M.
[Step S14]: M, generates additional portion _{P NEW} user profile from _{t M.} A specific example is shown in FIG.
[Step S15]: A new user profile _Pt is generated by synthesizing the existing user profile _Pt-1 and the additional portion _PNEW .

  The user profile includes, for example, a graph of words extracted from a message by morphological analysis and the message utterance time used at the time of synthesis.

  Step S14 (profile creation) of the statement content analysis unit 105 in FIG. 11 is performed as shown in FIG. 12, for example, and details thereof are as follows.

[Step S20]: the comment message M by morphological analysis assigns the result of the analysis in _{W M.}
[Step S21]: _W a complete graph that any word w with nodes belonging to _M is substituted into _{P NEW.}
[Step S22]: The message utterance time t _M is set as an attribute of P _NEW .

  Step S15 (profile synthesis) of the statement content analysis unit 105 in FIG. 11 is performed as shown in FIG. 13, for example, and details thereof are as follows.

[Step S30]: An arbitrary node of the existing user profile _Pt-1 is input to w.
[Step S31]: It is determined whether or not w is finished. If it is finished, the process is finished. If not, the process proceeds to step S32.
[Step S32]: It is determined whether any v belonging to W _M has a link to w. If the link is held, the process proceeds to step S33, and if not, the process proceeds to step S34.
[Step S33]: Create a link between v and w. Thereafter, the process proceeds to step S34.
[Step S34]: The next node of the existing user profile _Pt-1 is input to w, and the process returns to step S31 to repeat the processing.

  The task structure analysis unit 106 of the preprocessing unit 101 performs a process as shown in FIG. 14, for example, and details thereof are as follows.

[Step S40]: The new task table is substituted into TTAB (task map).
[Step S41]: The next task is extracted from the task map and assigned to T.
[Step S42]: It is determined whether or not T is completed. If it is completed, the process proceeds to step S46, and if not, the process proceeds to step S43.
[Step S43]: The next user of the task T is assigned to U.
[Step S44]: It is determined whether or not U is finished. If it is finished, the process returns to step S41 to repeat the process. Otherwise, the process proceeds to step S45.
[Step S45]: T and U are registered in TTAB (task map). After that, the process returns to step S43 to repeat the process.
[Step S46]: The time t _T when the task structure is updated is substituted into T. If it is sufficient to record the final task structure instead of recording the task structure at an arbitrary time, this step can be skipped.

  Next, the analysis unit 104 of the communication analysis system 100 will be described. The outline of the processing of the analysis unit 104 is as shown in FIG. 15, and details thereof are as follows.

[Step S50]: Substitute analysis conditions into C. The analysis conditions are input as shown in FIGS. To get the analysis target time from C to assign to t _A. The user of interest is identified from the analysis condition C and the user map UTAB and substituted for F. _A user file is searched with F and t _A and substituted into PF. When there are a plurality of profiles, one profile is specified by a desired rule, for example, the latest priority.
[Step S51]: Take out another user from the user map and substitute it into U.
[Step S52]: It is determined whether or not U is finished. If it is finished, the process ends. If not, the process proceeds to step S53.
[Step S53]: U, substituting the user profile hit by searching a user profile at _{t A} to _{P U.}
[Step S54]: A node list corresponding to P _U and C is taken out and substituted into W _U. How to extract the node list differs depending on the presence / absence of the analysis condition keyword, which will be described with reference to FIG.
[Step _S55]: W F, by calculating the similarity of _{W U} is substituted for _{S FU.} Various methods can be employed for calculation examples of similarity. For example, the method of Salton et al can be adopted (Salton, G, “Automatic Text Processing: The Transformation, Analysis, and Retrieval of Information by Computer”, Addison-Wesley, 1989). Alternatively, a word vector may be prepared for each user and the distance between the corresponding word vectors may be obtained.
[Step _S56]: is assigned to S by incorporating the aggregate _{S FU.}
[Step S57]: Output the total result of S.

  An example of step S54 (node list construction) of the analysis unit 104 is as follows.

[Step S60]: A keyword is input to KW. Empty if no keyword is specified.
[Step S61]: If KW is empty, the process proceeds to step S63, and if not, the process proceeds to step S62.
[Step S62]: assigns the W taken out neighbor nodes of KW in _{P U.} Proceed to step S64.
[Step S63]: assigns all nodes _{P U} to W. Proceed to step S64.
[Step S64]: W is output as a node list.

  Next, an operation example of the distance calculation unit 109 of the analysis unit 104 will be described. An example of the operation is as follows.

[Step S70]: Substitute analysis conditions into C. Assign the user of interest to F.
[Step S71]: Other users are extracted from the user map (UTAB) and assigned to U.
[Step S72]: It is determined whether or not U is finished. If it is finished, the process proceeds to step S75, and if not, the process proceeds to step S73.
[Step S73]: F, by calculating the distance between the U is substituted for _{D FU.} This distance is the minimum value (positive integer) of the number of tasks when tracing U from F or tracing F from U.
[Step _S74]: incorporated into aggregate _{D FU.} For example, it is represented by an array.
[Step S75]: The result of counting D is output.

  Next, an operation example of the similarity totaling unit 110 of the analysis unit 104 will be described. An example of the operation is as follows.

[Step S80]: Substitute analysis conditions into C. Substitute 1 for N.
[Step S81]: D corresponding to N is substituted into _DF .
[Step S82]: It is determined whether or not _DF is finished. If it is finished, the process is finished. If not, the process proceeds to step S83.
[Step S83]: _DF similarity is output.
[Step S84]: Count the number of users who have a certain degree of similarity.
[Step S85]: N is incremented by 1, and _DF is updated.

  Next, the analysis condition input unit 107 of the analysis unit 104 will be described. FIG. 20 shows an example of a graphical user interface for inputting analysis conditions. In this example, the target period start time and end time are input using a slide bar, a keyword is input as a character, and the user of interest is pulled down. However, the present invention is not limited to this. When the input of the analysis conditions is completed (such as pressing the “OK” button), each input is recorded in C and the process is terminated. When logarithmic r input is detected for each item, it is considered that analysis conditions have been input, and corresponding processing is performed. An example of this processing is as follows.

[Step S90]: The new analysis conditions are input to C.
[Step S91]: An input of analysis conditions is received from the user.
[Step S92]: Proceed to steps S93, S94, and S95 according to the analysis conditions.
[Step S93]: the time _{t A} is assigned to C. Proceed to step S96.
[Step S94]: The user of interest F is assigned to C. Proceed to step S96.
[Step S95]: The keyword KW is substituted for C.
[Step S96]: It is determined whether additional conditions remain. If it remains, the process returns to step S91 to repeat the process.
[Step S97]: Output analysis conditions.

  The analysis result output unit 111 of the analysis unit 104 outputs the analysis result (print, display). FIG. 21 shows an analysis result display example. In this example, the similarity is obtained from the user profiles at the start (the latest profile data time) and the end (the latest profile data) of the analysis target period, and the similarities are aggregated for each distance (average value processing). In this example, the slope becomes gentle (similarity decreases) with the passage of time, suggesting that the topic that is characteristic of the whole is decreasing.

  FIG. 22 shows another analysis result display example. In this example, the transition of a topic when a keyword is specified is shown. The size of the circle indicates the number of users with similar utterances. As shown in (A), when there is a sudden distribution of similar utterances at a distance, it suggests that an activity of another organization with some relationship (close problem awareness, etc.) has occurred, You can consider collaboration if necessary. You may make it display who corresponds to the corresponding intersection position (circle), and who belongs further, and which task it belongs to. (B) shows a case where the related utterance is in a decreasing tendency (gray circle) and a case where a certain amount of utterance is continuing (Mashinomaru). By grasping the decreasing trend of a specific topic, it is possible to discover the rigidity of the organization.

  In addition, this invention is not limited to the above-mentioned Example, A various change is possible. For example, as shown in FIG. 23, a dialog space to be analyzed (mailing list, electronic bulletin board, web access, etc.) is specified from the analysis condition input unit 107, and a dialog space structure analysis unit 120 (corresponding to the task structure analysis unit 106). Is used by switching the mailing list relationship extracting unit 121a, the electronic bulletin board relationship extracting unit 121b, and the like to calculate the distance between the user and the dialogue space based on the mailing list registration information, the electronic bulletin board registration information, and the like. Also good.

  FIG. 24 shows an example of processing when the interaction space structure analysis unit 120 uses the mailing list relationship extraction unit 121a. In this example, as shown in S107, the relationship extraction method is switched based on the support of the analysis condition input unit 107, and in this example, mailing list registration information is supplied. The processing in FIG. 24 is as follows.

[Step S100]: The new task table is substituted into TTAB (task map).
[Step S101]: The next task is extracted from the task map and assigned to T.
[Step S102]: It is determined whether or not T is finished. If finished, the process proceeds to step S46, and if not, the process proceeds to step S43.
[Step S103]: The next user of the task T is assigned to U.
[Step S104]: It is determined whether or not U is finished. If it is finished, the process returns to step S41 to repeat the process. If not, the process proceeds to step S45.
[Step S105]: T and U are registered in the TTAB (task map). After that, the process returns to step S43 to repeat the process.
[Step S106]: The time t _T when the task structure is updated is substituted into T. If it is sufficient to record the final task structure instead of recording the task structure at an arbitrary time, this step can be skipped.

  In addition to the communication analysis, the present invention can be configured as a tissue stiffness analysis system 400 that outputs an index of the degree of tissue stiffness. FIG. 25 shows an example of the configuration. In FIG. 25, portions corresponding to those in FIG. In this example, a stiffening determination unit 112 that performs statistical processing based on the similarity count result is provided, and the stiffening is digitized or thresholded to output the presence or absence of stiffening. For example, the stiffness may be determined based on the variance of the similarity of each distance before and after the target period, and various indicators such as determining that there is no rigidity when there are many similar utterance users at a large distance. Can be adopted.

It is a figure explaining the task (dialogue space) structure and distance of this invention. It is a figure explaining the relationship between the distance between tasks (dialogue space) of this invention, and the number of accessible users. It is a figure explaining the example of a display of the similarity distribution of this invention. It is a figure explaining the other example of a display of similarity distribution of this invention. It is a figure explaining the other example of a display of the similarity distribution of this invention. It is a figure explaining the example of the similarity distribution of this invention, and the content which it suggests. It is a figure which shows the structure of the Example of this invention as a whole. It is a figure which illustrates the example of the pre-processing phase of the above-mentioned Example typically. It is a figure which illustrates typically the example of the analysis phase of the above-mentioned Example. It is a figure explaining the example of the intermediate data used by the above-mentioned Example. It is a flowchart explaining the operation example of the statement content analysis part 105 of the pre-processing part 101 of the above-mentioned Example. It is a flowchart explaining the specific example of step S14 of the operation example of the statement content analysis part 105 of the pre-processing part 101 of the above-mentioned Example. It is a flowchart explaining the specific example of step S14 of the operation example of the statement content analysis part 105 of the pre-processing part 101 of the above-mentioned Example. It is a flowchart explaining the operation example of the task structure analysis part 106 of the pre-processing part 101 of the above-mentioned Example. It is a flowchart explaining the operation example of the profile similarity determination part 108 of the analysis part 104 of the above-mentioned Example. It is a flowchart explaining the specific example of step S54 of the operation example of the profile similarity determination part 108 of the analysis part 104 of the above-mentioned Example. It is a flowchart explaining the operation example of the distance calculation part 109 of the analysis part 104 of the above-mentioned Example. It is a flowchart explaining the operation example of the similarity totaling part 110 of the analysis part 104 of the above-mentioned Example. It is a flowchart explaining the operation example of the analysis condition input part 107 of the analysis part 104 of the above-mentioned Example. It is a figure explaining the example of the graphical user interface of the analysis condition input part 107 of the analysis part 104 of the above-mentioned Example. It is a figure explaining the example of a display of the analysis result output part 111 of the analysis part 104 of the above-mentioned Example. It is a figure explaining the other example of a display of the analysis result output part 111 of the analysis part 104 of the above-mentioned Example. It is a block diagram explaining the modification of the above-mentioned Example. It is a flowchart explaining the modification of the said Example. It is a block diagram explaining the other modification of the said Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Communication analysis system 101 Preprocessing part 102 User profile database 103 Task structure database 104 Analysis part 105 Statement content analysis part 106 Task structure analysis part 107 Analysis condition input part 108 Profile similarity determination part 109 Distance calculation part 110 Similarity totalization part 111 Analysis result output unit 112 Stiffening determination unit 120 Dialogue space structure analysis unit 121a Relationship extraction unit for mailing list 121b Relationship extraction unit for electronic bulletin board 200 Communication system 201 Communication system log 300 Computer 301 Program 400 Tissue hardening analysis system

Claims (16)

Similarity determination means for determining the similarity between communication contents of members of the organization and communication contents of other members of the organization;
An inter-member distance calculating means for calculating a distance between the members of the organization and other members;
A communication analysis apparatus comprising: a totaling unit that totalizes the similarity of the similarity determination unit based on the distance calculated by the inter-member distance calculation unit.   The communication analysis apparatus according to claim 1, wherein the similarity determination unit determines a similarity between the communication contents based on words included in the communication contents.   The communication analysis apparatus according to claim 1, wherein the inter-member distance calculation means calculates the inter-member distance based on an organization part to which the member belongs.   4. The distance from a target member to another member is obtained by connecting the two members belonging to each tissue part and determining the minimum number of tissue parts on the route reaching the other member from the target member. Communication analysis equipment.   5. The communication analysis apparatus according to claim 3, wherein the organization part is a group that can be defined by an electronic communication tool.   The communication analysis apparatus according to claim 1, wherein the counting unit totals the number of similarities whose similarity exceeds a predetermined threshold.   The communication analysis apparatus according to claim 1, further comprising a display unit that displays a totaling result of the totaling unit.   The communication analysis apparatus according to claim 1, further comprising a determination unit that determines whether the tissue is rigid based on a counting result of the counting unit.   The communication analysis apparatus according to claim 1, further comprising an input unit configured to input member information serving as a distance reference.   The communication analysis apparatus according to claim 9, wherein the input unit inputs information specifying a time range of communication to be determined.   The communication analysis apparatus according to claim 9 or 10, wherein the input means inputs a keyword for narrowing down communication. Similarity determination means for determining the similarity between communication contents of members of the organization and communication contents of other members of the organization;
An inter-member distance calculating means for calculating a distance between an organization member and another member based on a group in which the member is registered;
A communication analysis apparatus comprising: a totaling unit that totalizes the similarity of the similarity determination unit based on the distance calculated by the inter-member distance calculation unit.   The communication analysis apparatus according to claim 12, wherein the group is a group that can be defined by an electronic communication tool. Similarity determination means for determining the similarity between communication contents of members of the organization and communication contents of other members of the organization;
An inter-member distance calculating means for calculating a distance between the members of the organization and other members;
A counting means for counting the similarity of the similarity determination means based on the distance calculated by the distance calculation means between members;
Display means for displaying the counting result of the counting means;
A tissue rigidity analysis apparatus comprising: a determination unit that determines whether the tissue is rigid based on a totaling result of the totaling unit. A step of determining a similarity between communication contents of members of the organization and communication contents of other members of the organization;
A member-to-member distance calculating means calculating a distance between the members of the organization and the other members;
A tally means for tallying the similarity of the degree of similarity determination means on the basis of the distance calculated by the distance calculation means between members. Similarity determination means for determining the similarity between communication contents of members of the organization and communication contents of other members of the organization;
An inter-member distance calculating means for calculating a distance between the members of the organization and other members;
A computer program for communication analysis, which is executed by a computer in order to realize a summing means for summing up the similarities of the similarity determining means on the basis of a distance calculated by the distance calculation means between members.