JP2007018201A - Institution communication analyzing device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To extract the deadlock state of a communication network regardless of the timing for collecting information relating to a network. <P>SOLUTION: A communication history storing part 12 stores a message log. An analytic history extracting part 11 extracts a history as the object of analysis. A group information storing part 14 stores information for converting a history by individual such as members belonging to a group. An individual history summarizing part 13 converts the target history by individual, and summarizes them. An evaluation period determining part 15 calculates an inspection period for deciding change in a comparison reference period based on the history data. A verification part 16 statistically verifies whether the comparison reference period is significantly different from the comparison target period. An output part/recording part 17 outputs or records the final result. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an organization communication analysis technique for analyzing organization communication and determining organization rigidity.

  In general, in the case of research projects and development projects, it has been verified that the amount of communication outside the organization decreases, and the performance of the technical service project decreases as the amount of communication within the organization decreases, and is pointed out as NIH syndrome (Not Invented Here) (Katz, Ralph and Thomas J. Allen, R & D Management, 1982, 12, pp. 7-19). There is a relationship that the project performance is high from 1.5 to 4.9 years after the project, and the decrease in the frequency of communication is correlated with the decrease in performance thereafter.

  It is desired to focus on communication within the organization to determine the degradation of the organization's performance, typically the organization's rigidity, and to provide it as information for organizational management.

  By the way, from the viewpoint of organizational communication, various methods for visualizing and indexing the relationship information between speakers by analyzing message logs in e-mails and electronic bulletin boards have been devised.

  As a technique for handling a message log by paying attention to a network structure of communication, for example, Patent Document 1 discloses a method of analyzing a message log and sharing related information. Patent Document 2 discloses a method of analyzing message logs and displaying related information in an easy-to-understand manner.

  Therefore, if such communication analysis technology can automatically detect from the message log whether or not the communication network structure has changed, such as when new communication is no longer born and communication is stagnant, organizational performance will decline. It is thought that can be stopped.

  However, the network structure is almost sparse at a certain moment, and a time span such as week or month is required. For this reason, there is a problem that the result depends on how to determine the aggregation period and the aggregation start timing. For example, FIG. 1 shows the occurrence of communication between the members A to C as vertical line segments. In this example, the period A and the period A ′ are only slightly shifted in timing with the same time width. Despite being, the network structure looks different. That is, in the period A, (B → A, B → C), while in the period A ′, (B → A, B → C, D → C).

It is desirable to be able to appropriately analyze the communication structure and the like without being significantly affected by the counting period and its start timing.
JP-A-10-301905 JP 2003-85570 A

  The present invention has been made in consideration of the above circumstances, and an object thereof is to provide a communication analysis technique that can extract a stuck state of a communication network regardless of a network information or other information aggregation timing. .

  According to this invention, in order to achieve the above-mentioned object, the configuration as described in the claims is adopted. Here, prior to describing the invention in detail, supplementary explanations of the claims will be given.

  That is, according to one aspect of the present invention, in order to achieve the above-mentioned object, in the organization communication analyzer: communication log storage means for recording a communication log in the organization; A reference period pattern totaling unit that extracts a communication log from the communication log storage unit and totals a communication pattern; and extracts a communication log within the organization in a predetermined examination period from the communication log storage unit to obtain a communication pattern. An inspection period pattern totaling means for summing; and a fitness level determining means for determining the fitness of classification probability of the totaling of the reference period pattern totaling means and the totaling time of the test period pattern totaling means; And so as to determine the stiffness of the tissue.

  In this configuration, the communication patterns are aggregated for each of the reference period and the inspection period, the classification probabilities for the aggregation of each communication pattern are determined, and based on this, the fitness between each period is determined. Thus, it is possible to determine the degree of communication pattern squeezing and thus the rigidity of the tissue. The classification probability means the probability of classification into a communication pattern. The probability may be expressed by a probabilistic numerical value, may be a numerical value between 0 and 1, or may not be. It may be a numerical value normalized for each comparison target.

  The preferred reference period depends on the organization's stiffening process, but is typically half a year considering that the peak performance of a normal project is a span of 1.5 to 4.9 years However, it is not limited to this.

  The inspection period is a period sufficient to obtain sufficient accuracy, and is a period sufficient for the smallest number of communication patterns to be ten. The inspection period typically goes back from the present time, but if the accuracy is not affected, the inspection period may end at a time slightly back from the current time.

  The inspection period may overlap the reference period, or the reference period may not overlap the inspection period. That is, the inspection period may be set in a part of the reference period recently, or the inspection period may be set after the reference period.

  Any communication within the organization can be used as long as it can be maintained and managed in an organizational unit. Typically, it is the entire company, a department within the company, etc., but is not limited to this. An electronic bulletin board unit (and its thread) and a mailing list unit to be exchanged are also “organizations”.

  The communication pattern is typically defined by the communication source and destination. Communication contents such as keywords may be used as communication patterns. The number of communications may be the distribution of communications over time.

  The reference period may be specified by default, or means for specifying the reference period may be provided.

  In addition, an inspection period determination unit may be provided that determines the start of the inspection period as a time point at which a predetermined number of communication patterns having the minimum frequency can be sampled.

  The fitness level determination means typically determines the fitness level of each category of the total classification based on the chi-square distribution.

  Such communications typically include electronic communications, including email, instant messaging, posting on electronic bulletin boards, web access and electronic document access, telephone calls, verbal conversations and access between parties. It is not limited to. The approach between the parties can be detected by employing an RFID tag or the like.

  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 the present invention, it is possible to extract the stalemate status of a communication network regardless of the aggregation timing of information about the network such as the network structure.

  Hereinafter, a communication analyzer according to an embodiment of the present invention will be described.

  This embodiment is typically realized as software on a computer. The functional blocks described below are implemented by cooperating computer hardware resources and software resources.

  In this embodiment, attention is paid to the occurrence rate of communication from one member to another using the concept of queuing theory. When the incidence of communication is low compared to the analysis period, when comparing one period with another period to compare changes, the difference between the occurrence period and the other period is 1 It looks big at 0. Therefore, by extracting only communication links with a sufficient number of occurrences within the target period and statistically testing whether the number of occurrences has changed, it is derived whether there is a change in the main part of the network structure.

  Since there is a change in the long term, focus on recent logs that are stiffening. Since NIH says that there is a performance peak from 1.5 to 4.9 years, the extraction of stiffening should be seen from the last half year to about one year as the time range. This is called an analysis target period. Hereinafter, it may be referred to as a reference period or a comparison reference period in comparison with the test object. Here, for example, in a half year, a pattern with many utterances (20 utterances or more) is taken out. Among them, the period in which the last ten utterances of the smallest pattern are expected is tested, and it is examined whether there is a change with respect to the overall ratio (the ten utterances are used as a reference to approximate accuracy to the χ-square distribution) To make it higher). If no change is observed, it is determined that there is a high possibility of being rigid. It should be noted that the test target period may be referred to as a comparison target period, an inspection period, or a test period.

  FIG. 2 shows the relationship between the comparison reference period and the comparison target period. In this example, communication of B → A, D → C, and B → C is shown, and D → C having a small frequency (number of cases) is shown. Ignored and targets B → A and B → C with a high frequency, and the communication of B → A with a low frequency is traced back by 10 utterances, and the time point is set as the start point of the comparison target period. The starting point of the comparison reference period is, for example, a time point that goes back half a year from the present time. In this example, the comparison reference period and the comparison target period overlap, but may be separated as indicated by the broken line. It is determined whether there is a high possibility that the target pattern has become rigid using a deviation between the comparison reference period and the comparison target period.

  FIG. 3 schematically shows the configuration of the embodiment. As described above, the functional blocks constituting the communication analysis apparatus according to the embodiment are realized by installing the program 301 in the computer 300.

  In FIG. 3, the communication analysis apparatus 100 includes an analysis target instruction unit 10, an analysis history extraction unit 11, a communication history holding unit 12, an individual history totaling unit 13, a group information holding unit 14, an evaluation period determining unit 15, a testing unit 16, An output unit / recording unit 17 and the like are included. A communication system 200 is provided in association with the communication analyzer 100. The communication system 200 is an apparatus for exchanging electronic messages such as electronic mail. Below, it demonstrates as a server which provides an electronic mail service. When the history holding unit of the communication system 200 can be used, it may be used for the communication history holding unit 12.

  The communication history holding unit 12 holds a history such as a message log (electronic mail log). A history may be extracted from the network. The analysis history extraction unit 11 extracts a message log from the communication history holding unit 12 regarding the target period and the target group. The analysis target instruction unit 10 specifies a target period, a target member, a target group, and the like. From the NIH theory, it is appropriate to use history information from the last half year to about one year as the target period.

  The group information holding unit 14 holds information that can convert the history into individual units, such as members belonging to the group. When the target member is designated from the outside, it can be realized without having a configuration. The individual history totaling unit 13 converts the target history into individual units and totals them. The evaluation period determination unit 15 calculates a comparison target period for determining a change in the comparison reference period (for example, half a year) based on the history data. The test unit 16 statistically tests whether there is a significant difference between the comparison reference period and the comparison target period. The output unit / recording unit 17 outputs or records the final result.

  FIG. 4 shows the operation of the communication analysis apparatus according to this embodiment. In this figure, the analysis history extraction unit 11 extracts message logs for a target period, that is, six months to one year, from the communication history holding unit 12. (S10). When specifying the group to be analyzed, the necessary message logs are also narrowed down here. The data structure of a communication log (typically a mail log, hereinafter simply referred to as a mail log) is, for example, as shown in FIG. An example of the mail log is as shown in FIG.

  Thereafter, when there is a message log addressed to the group in the message log, the individual history totaling unit 13 refers to the group member list of the group information holding unit 14 and converts it into individual units, and further, the sender and destination The number of cases is totaled for each combination (link) and sorted in descending order (S11). The group member list is, for example, as shown in FIG. The relationship between the member's mail address and the user ID is held as shown in FIG. 8, for example. FIG. 9 shows an example in which the mail log transmission / reception relationship is converted into member units.

  After that, in the target link extraction, only 20 or more messages are extracted from the total result of the combination of sender and destination (link) (S12). An example of this is shown in FIG.

  The evaluation period determination unit 15 extracts the last 10 cases for the smallest combination of senders and destinations (links), and the first outgoing call among them. Is obtained (S13). FIG. 11 shows an example of the last ten extraction examples and corresponding time examples. After this time, it becomes the subject of examination. An example of the test target is shown in FIG.

  The inspection unit 17 performs a significant difference test (S14). That is, for each of the other links, the number of cases after the obtained time is totaled, and a significant difference in the number of cases between the target period and the obtained time is determined by a statistical test. The output unit / recording unit 17 outputs or records the result (S14).

  FIG. 13 shows the number of target communication patterns in the comparison reference period and the comparison target period, and shows the expected frequency of the target communication pattern in the comparison target period based on the number of target communication patterns in the comparison reference period. Yes. The comparison target period is about 0.46 times the comparison reference period, and an expected frequency is obtained by multiplying the frequency of the comparison reference period by this.

  In this example, as shown in FIG. 13, the chi-square value is 50.57993, the degree of freedom n (number of target patterns) is 21, and the chi-square distribution value 31.41042 when the significance level is 5%. large. Therefore, it cannot be said that the number of comparison target periods is equal to the comparison standard. That is, since a change is recognized in the structure of the organization communication network, it is determined that the organization communication is not rigid. The testing unit 16 performs testing by holding the χ square distribution value as a table, for example.

  According to this embodiment, attention is paid to a typical communication pattern, and the comparison target period is set so that a sufficient number of cases can be sampled, so that the degree of communication sticking can be reliably determined.

  Next, a modification of this embodiment will be described.

  From the access log in the document sharing tool, it is also possible to examine the communication structure that mediates the document. Here, in the individual history totaling process in the individual history totaling unit 13, the verification can be performed in the same manner by interpreting the referrer who has accessed the document as the destination and the registrant of the document as the sender. In this case, the access time is not the document registration time corresponding to the transmission time, but the access time that can be interpreted as having actually received the information.

  FIG. 14 shows an example of an access log database. FIG. 15 shows a log extracted in document units and an example of converting it into a link. FIG. 16 shows an example in which the number of cases is totaled for each document registrant and reference person, and is sorted in descending order of the number of cases (20 cases or less are omitted). The combination of the minimum number of evaluation targets (4, 6) is shown in FIG. The number of cases after the 10th time from the back (comparison target) is counted and tested.

  In this modification, the chi-square value is 0.700206, the chi-square distribution value is 22.36203 (degree of freedom n = 13, significance level = 5%), and the probability of the comparison target period with respect to the comparison criterion Can be determined to be equal. That is, it is determined that there is a possibility that the communication of the organization has become rigid because no change is recognized in the structure of the communication network of the organization.

  Next, another modification will be described.

  It is also possible to examine the structure of communication by telephone call from a private branch exchange (PBX) telephone call record. In the case of a telephone, there is a distinction as to which was made, but there is no distinction between the transmission direction such as the sender and the destination. Therefore, in the individual history tabulation process in the individual history tabulation unit 13, the number of cases is determined without distinguishing the link direction Are added and tested in the same manner. In this case, in order to improve the verification accuracy, it is preferable to extract the top calls by the number of cases, and the comparison for the verification is performed by the total call time. This is for the purpose of taking into account the fact that the amount of information transmitted between a long-time communication and a short-time communication in a single call is different. In addition, the call start time is used as the break time.

  FIG. 17 shows an example of the total result. In this example, the χ square value is 2.742424, the χ square distribution value is 15.50731 (degree of freedom n = 8, significance level = 5%), and comparison is made. The ratio of the call time in the comparison target period is equal to the reference. That is, it is determined that there is a possibility that the communication of the organization has become rigid because no change is recognized in the structure of the communication network of the organization.

  The present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the invention. For example, in the above-described example, the chi-square test is performed by paying attention to the appearance rate in the multinomial distribution, but various test methods such as a t distribution test based on the average value can be adopted. In the above example, e-mail is mainly described. However, the present invention can also be applied to communication in instant message, electronic bulletin board, or web access. In addition, it detects the location of a member using a position detection system using an RFID (Radio Frequency Identification) tag, an infrared badge, etc., and detects communication within the organization by determining that there is communication when in the same location. Then, the communication stuck state may be determined based on the pattern.

It is a figure explaining the aspect in which communication is exchanged. It is a figure explaining the comparison reference period and comparison object period of the Example of this invention. It is a block diagram which illustrates typically the composition of the above-mentioned example. It is a flowchart explaining the operation example of the said Example. It is a figure explaining the example of the data structure of the mail log of the above-mentioned Example. It is a figure explaining the example of the mail log of the above-mentioned Example. It is a figure explaining the example of the group member list of the above-mentioned Example. It is a figure explaining the example of the mail address of the member of the above-mentioned Example. It is a figure explaining the example which converts the mail log of the said Example into a member unit. It is a figure explaining the example of selection of the target pattern of the above-mentioned example. It is a figure explaining the example of determination of the start of the comparison object period of the above-mentioned Example. It is a figure explaining the example of the total result of the comparison reference period and comparison object period of the above-mentioned Example. It is a figure explaining the example of the test result of the above-mentioned Example. It is a figure explaining the access log of the modification (modification 1) of the above-mentioned Example. It is a figure explaining the example of the test result of the modification of the above-mentioned Example. It is a figure explaining the example of the test result of the other modification (modification 2) of the above-mentioned Example. It is a figure explaining the example of the test result of the other modification of the above-mentioned Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Analysis object instruction | indication part 11 Analysis history extraction part 12 Communication history holding part 13 Individual history totaling part 14 Group information holding part 15 Evaluation period determination part 16 Test part 17 Output part / Recording part 100 program

Claims (9)

A communication log storage means for recording a communication log in the organization;
A reference period pattern totaling unit that extracts a communication log in the organization in a predetermined reference period from the communication log storage unit and totals a communication pattern;
Inspection period pattern totaling means for extracting communication logs within the organization in a predetermined inspection period from the communication log storage means and totalizing communication patterns;
A fitness level determining means for determining the level of fitness of the classification probability of the totaling of the reference period pattern totaling means and the totaling time of the inspection period pattern totaling means,
A tissue communication analysis apparatus characterized by determining the rigidity of the tissue based on the fitness.   The organizational communication analysis apparatus according to claim 1, wherein the communication pattern is defined by a communication source and destination.   The organization communication analysis apparatus according to claim 1 or 2, further comprising means for designating the reference period.   The organization communication analysis apparatus according to any one of claims 1 to 3, further comprising an examination period determining unit that determines an initial period of the examination period as a time point at which a predetermined number of communication patterns having a minimum frequency can be sampled.   The organizational communication analysis apparatus according to any one of claims 1 to 4, wherein the fitness determination unit determines the fitness of the classification probability of aggregation based on a χ square distribution.   6. The organizational communication analysis apparatus according to claim 1, wherein the communication is electronic communication including electronic mail, instant message, posting on an electronic bulletin board, web access, and electronic document access.   The organizational communication analysis device according to claim 1, wherein the communication includes a telephone call, a verbal conversation, and an approach. A communication log storage means for recording a communication log in the organization;
A step in which a reference period pattern totaling unit extracts a communication log in the organization in a predetermined reference period from the communication log storage unit and totals a communication pattern;
A step in which an inspection period pattern totaling unit retrieves a communication log in the organization in a predetermined inspection period from the communication log storage unit and totals a communication pattern;
And a step of determining a fitness of the classification probability of the aggregation of the reference period pattern totaling means and the totalization of the inspection period pattern totaling means.
A tissue communication analysis method, comprising: determining the rigidity of the tissue based on the fitness. A communication log storage means for recording a communication log in the organization;
A reference period pattern totaling unit that extracts a communication log in the organization in a predetermined reference period from the communication log storage unit and totals a communication pattern;
Inspection period pattern totaling means for extracting communication logs within the organization in a predetermined inspection period from the communication log storage means and totalizing communication patterns;
Executed by a computer in order to realize the suitability determination means for determining the suitability of the classification probability of the summation of the reference period pattern summing means and the summation of the inspection period pattern summarization means,
A computer program for tissue communication analysis, wherein the tissue stiffness is determined based on the fitness.