Information processing device, information processing method, and program

Abstract

The information processing device comprises: an information acquisition unit that acquires, in response to a search request from a first user asking to search for a candidate user from among a plurality of users, first text information which is handled in a terminal device used by the first user; a condition generation unit that classifies, in accordance with the frequency of appearance in second text information pertaining to previous work of the first user, words contained in the first text information into a first word group having a frequency of appearance of lower than a first threshold and a second word group having a frequency of appearance of not lower than the first threshold; a search processing unit that extracts a candidate user from among the plurality of users in accordance with the frequency of appearance of words belonging to the first word group in third text information pertaining to previous work of each of the users, identifies manual types that each candidate user has browsed in the past, and calculates a priority level for each candidate user such that a higher priority level is given for greater frequency of appearance of words belonging to the first word group in the manuals; and a result output unit that outputs information for presenting, to the first user, information pertaining to the candidate user and including the priority level.

Description

[Technical field]

[0001] An embodiment of the present invention relates to an information processing device, an information processing method, and a program. [Background technology]

[0002] In companies, etc., tasks are assigned to business practitioners (human resources, accounting, etc.), and business is carried out according to the assigned task. Business practitioners in each task are human resources with a fairly uniform skill set, including experience. Even if they are not directly involved, each business practitioner can naturally hear conversations between people around them, and is considered to be able to sense on a daily basis who is consulting whom about what, who is knowledgeable about what, and who is having trouble with what.

[0003] In the future, as work patterns become more diverse, the skill sets of each employee are expected to become more diverse. In addition, it is expected that work will be carried out in a mix of remote and office environments, and opportunities to hear the above-mentioned interactions will decrease.

[0004] Therefore, it is expected that the situations in which each practitioner needs help will be diverse. For those who need help, it may become more difficult than ever to understand who is knowledgeable about what, as a prerequisite, and to find the right person to consult when a problem occurs. Furthermore, it is expected that it will become more difficult for those who want to help to determine which problems they can easily help with based on their own experience.

[0005] In this way, when a problem occurs in business that requires consulting with someone, there is a need to search for a person who has experience with the problem. A technology has been proposed that enables a search for users in a company who have a wealth of knowledge about the topic to be learned (for example, see Non-Patent Document 1).

[0006] The method proposed in Non-Patent Document 1 derives information about the problem of the person who is the business operator when the problem occurs from the "current business situation" and "accumulated business experience" expressed as vocabulary distribution in the terminal log, and extracts consultation candidates who are likely to have the business experience to solve the problem under the same situation by comparing the vocabulary distribution.The method proposed in Non-Patent Document 1 then selects as the consultant, from among these consultation candidates, someone who has more business experience similar to that of the person at the time of the consultation and who is likely to be able to smoothly understand the problem and present a solution. [Prior art documents] [Non-patent literature]

[0007] [Non-Patent Document 1] Pairing consultants by matching work experience, R6-9, Tsukuba Forum 2020 ONLINE, October 29, 2020 - October 30, 2020. Summary of the Invention [Problem to be solved by the invention]

[0008] However, there are cases where a person cannot solve a problem simply because he or she has a lot of experience solving problems in the same circumstances as the person seeking advice.

[0009] The present invention has been made in light of the above-mentioned circumstances, and has an object to provide a technique that enables a user to more appropriately search for a place to seek advice on a problem. [Means for solving the problem]

[0010] In order to solve the above problems, a first aspect of the present invention is an information processing device, an information acquisition unit that acquires first text information handled in a terminal device used by a first user in response to a search request issued by the first user to search for a candidate user from among a plurality of users; a condition generating unit that classifies words included in the first text information into a first group of words whose occurrence frequency is lower than a first threshold and a second group of words whose occurrence frequency is equal to or higher than the first threshold, according to the occurrence frequency in second text information related to a past work of the first user; extracting candidate users from among the plurality of users according to the frequency of appearance of words belonging to the first word group in third text information related to past work of each user; determining manuals that each of the extracted candidate users has previously referred to, and calculating a priority of each candidate user such that the higher the frequency of appearance of a word belonging to the first word group in the manuals, the higher the priority; A search processing unit; a result output unit that outputs information for presenting information about the candidate users, including the priority levels, to a first user; The present invention is designed to provide the above. Effect of the Invention

[0011] According to a first aspect of the present invention, in response to a search request from a first user, text information handled by a terminal device used by the first user is acquired. Words included in the text information are classified according to their frequency of appearance in second text information related to the first user's past work. Words that appear less frequently in the second text information related to the first user's past work can be rephrased as words with which the first user has little experience, or words about which the first user needs information. For such words that appear less frequently in the second text information related to the first user's past work among the words in the text information handled by the first user's terminal device, candidate users are extracted based on their frequency of appearance in the second text information related to the search target user's past work. Then, manuals that each of the extracted candidate users has referred to in the past are determined, and the priority of each candidate user is calculated so that the higher the frequency of appearance of words belonging to the first word group in those manuals, the higher the priority.

[0012] In this way, the candidate users are extracted according to the frequency of occurrence of the necessary information in the third text information related to the past work of the search target user, so there is no need to tag the search target user in advance. In addition, since the search is performed based on information that is estimated to be useful to the first user, taking into account the experience of the first user, more appropriate search results can be obtained. Furthermore, since there is no need to input a search query, it is convenient for the first user and a search that flexibly reflects the current situation can be performed without being bound by a specific keyword. Then, the priority of each candidate user is calculated so that the higher the frequency of occurrence of words belonging to the first word group in the manuals that the extracted candidate user has referred to in the past, the higher the priority is, so that users who have versatile knowledge (the meaning and reason of each work) or have come into contact with general information can be preferentially selected as consultants, which increases the possibility of solving the problem.

[0013] That is, according to the first aspect, it is possible to provide a technique that enables a more appropriate search for a place to seek advice on a problem. [Brief description of the drawings]

[0014] [Figure 1] FIG. 1 is a diagram showing an example of the overall configuration of a search system including a search server as an information processing device according to a first embodiment of the present invention. [Diagram 2] FIG. 2 is a schematic diagram showing an overview of the process of the search server shown in FIG. [Diagram 3] FIG. 3 is a schematic diagram showing an outline of the process up to the extraction of consultation candidates with recommendation levels shown in FIG. [Figure 4] FIG. 4 is a block diagram showing an example of a functional configuration of the search server serving as the information processing device according to the first embodiment. [Diagram 5] FIG. 5 is a block diagram showing an example of a hardware configuration of the search server shown in FIG. [Figure 6] FIG. 6 is a flowchart showing an outline of the process performed by the search server shown in FIG. [Figure 7] FIG. 7 is a flowchart illustrating an example of a procedure for a terminal handling content grasping process from the processes illustrated in FIG. [Figure 8] FIG. 8 is a diagram showing an example of text information acquired by the process shown in FIG. [Figure 9] FIG. 9 is a diagram illustrating an example of data stored in the process illustrated in FIG. [Figure 10] FIG. 10 is a flowchart illustrating an example of a procedure of the pairing condition generating process from the processes illustrated in FIG. [Figure 11] FIG. 11 is a schematic diagram showing a first example of the determination process of the processes shown in FIG. [Figure 12] FIG. 12 is a schematic diagram showing a second example of the determination process of the processes shown in FIG. [Figure 13] FIG. 13 is a diagram showing an example of data stored in the process shown in FIG. [Figure 14] FIG. 14 is a flowchart illustrating an example of a procedure of the pairing process from among the processes illustrated in FIG. [Figure 15] FIG. 15 is a flowchart showing an example of a procedure for counting the number of times the work implementers referred to the manuals shown in FIG. [Figure 16] FIG. 16 is a diagram showing an example of data stored in the process shown in FIG. [Figure 17] FIG. 17 is a flowchart illustrating an example of a procedure of a pairing result output process from the processes illustrated in FIG. [Figure 18] FIG. 18 is a schematic diagram showing an overview of the process of a search server serving as an information processing device according to the second embodiment of the present invention. [Figure 19] FIG. 19 is a flowchart showing an example of a procedure of the pairing process in the second embodiment. [Figure 20] FIG. 20 is a flowchart illustrating an example of a procedure for the consultation candidate extraction process illustrated in FIG. [Figure 21]FIG. 21 is a flowchart illustrating an example of a procedure for the consultation candidate narrowing down process illustrated in FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the following, elements that are the same as or similar to elements already described will be given the same or similar reference numerals, and duplicated descriptions will basically be omitted. For example, when there are multiple identical or similar elements, a common reference numeral may be used to describe each element without distinguishing between them, and a subnumber may be used in addition to the common reference numeral to describe each element with distinction between them.

[0016] [First embodiment] (1) Composition (1-1) Search system 1 is a diagram showing an example of the overall configuration of a search system 100 including an information processing device according to a first embodiment of the present invention. The search system 100 includes a search server 1 and user terminals UT1, UT2, ..., UTn (hereinafter collectively referred to as "user terminals UT") that can communicate with each other via a network NW.

[0017] The information processing device according to the first embodiment can be implemented as a search server 1. The search server 1 is, for example, a server computer or a personal computer managed by an organization such as a company, a school, an association, or a membership community.

[0018] The search server 1 receives requests from users who have encountered a problem (hereinafter referred to as "consultant" or "consultation user") by communicating with each user terminal UT, searches for candidates for users who will receive consultation (hereinafter referred to as "consultee" or "consultation user"), and presents the search results to the consultation user. The search server 1 can also notify the consultation user that they have been extracted as a candidate, that a problem has occurred, and information about the consultation user. In this way, the search server 1 recommends candidate consultation users to the consultation user and makes the candidate consultation users aware of the occurrence of a problem, thereby supporting the connection between the consultation user and the consultation user (herein referred to as "pairing").

[0019] The network NW may be, for example, the Internet, an intranet, a local area network (LAN), a wide area network (WAN), etc., and may include a wired or wireless network.

[0020] The user terminal UT is an information processing terminal used by a user, and may be, for example, a personal computer, a smartphone, a tablet terminal, a mobile phone, etc. A user belongs to an organization such as the one described above that manages the search server 1, and performs business operations using the user terminal UT, and can also communicate with other users via the user terminal UT. The user terminal UT may be a dedicated terminal for each user, or may be shared by multiple users.

[0021] In the following, as an example, a situation will be described in which users as business practitioners belonging to a company use their respective user terminals UT to use the search system 100. Note that the search server 1 may be a device separate from the user terminals UT, or may be integrated with any of the user terminals UT.

[0022] As described above, in today's business environment, the skill sets possessed by each user are becoming more diverse, and it is becoming more difficult to naturally perceive on a daily basis "who is knowledgeable about what" and "who is having trouble with what." Therefore, when a problem occurs that requires the consultation of others, the user who wants to consult needs to search for a person who has experience with the problem. The search server 1 is intended to enable both the person seeking advice and the person seeking advice to find a person who can share "the business situation at the time the problem occurred" and "a solution appropriate to the situation" with minimal effort (including explanation) when such a problem occurs.

[0023] In conventional methods, content is stored that has a mechanism for explicitly retaining the creators of created / registered documents (emails, web pages, office documents, etc.), and a person wishing to find an expert enters a search query or keyword to extract the creators / users of content that contains that keyword. To perform a search, it is necessary to actively tag each work experience to the relevant person (for example, "#reception", "#development", "#budget management") in order to clarify on the system who has what knowledge. Alternatively, a mechanism is needed to automatically tag each work experience to the relevant person, and technology for this is required.

[0024] In addition, in the conventional method, when extracting consultation candidates using keywords, the contents of documents created / registered by the candidate and the contents of emails sent / received are used as information on the experience of the consultation candidate. However, in this case, the holders of experience related to a certain keyword are limited to the creators / registerers of documents, in addition to the sent / received emails.

[0025] For example, if a problem occurs when using a business system, documents containing the words "related to problems (such as the contents displayed on the business system screen) or business status" used as search keywords are mainly those created / registered by the business system provider, such as the business system user manual, and there are few documents created / registered by the user. Therefore, the business system provider is more likely to be extracted as a person to consult than the user of the business system. However, the business system provider does not usually perform the work of a user. In addition, even if a problem occurs when using a business system, it may not be related to the method of using the business system, but may be related to a specific business content that is being performed on the business system that is unknown to the business system provider. For these reasons, the business system provider is unlikely to have experience in solving problems, and rather the user is more likely to meet the requirements of the case. As a result, there is a risk that people who actually meet the requirements of the case will be overlooked as candidates to consult.

[0026] Furthermore, the conventional method extracts consultation destinations based only on the keywords entered by the person seeking advice. The search keywords, for example, as shown in (Example a) to (Example c) below, do not distinguish between those related to business situations and those related to problems. (Example a) When the search keywords selected by the person seeking advice are limited to vocabulary related to the work situation: Even if it is possible to extract consultation partners who have ample experience in the same work situation as the person seeking advice, it is not taken into consideration whether they have experience in solving problems. Therefore, there is a high possibility that the consultation partner does not have experience in solving problems, and in that case, it will not be useful in solving the problem. (Example b) When the search keywords selected by the person seeking advice are limited to vocabulary related to the problem: Even if it is possible to extract a consultation partner with ample experience in the problem, it is not taken into consideration whether they have experience in the same work situation as the person seeking advice is facing the problem. Therefore, there is a high possibility that the consultation partner does not have experience in the work situation, and in that case, the person seeking advice will need to explain a lot to the consultation partner in order to have them understand the work situation at the time the problem occurred. Even if the consultation partner shares their own experience as it is, it will not be a solution that is suited to the person seeking advice's work situation, and they will need to prepare a new explanation that is suited to the person seeking advice's work situation. (Example c) When the search keywords selected by the person seeking advice contain vocabulary related to both the work situation and the problem: There is no distinction between whether the search keywords are related to the work situation or the problem, and for example, the creator / user of the content that contains more search keywords will be extracted as the person to consult. Therefore, the person to consult that is extracted will not necessarily be the creator / user of the content that contains vocabulary related to the problem, and the same problem as above (Example a) will occur.

[0027] In consideration of the above problems, the search server 1 or the search system 100 according to the first embodiment focuses on at least one of the following points. [Point 1] Text information handled by each user terminal UT is acquired in the background, recorded and stored in association with the business performer (user), and used to understand the business experience of each business performer. [Point 2] Business information, which is the information the person seeking advice is currently dealing with (at the time of consultation), is divided into "problem information," which is information about the problem the person seeking advice wants to discuss, and "peripheral information," which is information about the work the person seeking advice is performing. [Point 3] Candidates for consultation are selected based on their work experience related to problem information, and a higher recommendation level is assigned to those who have more information on both the problem and related issues. [Point 4] Among those who are assigned a high recommendation level, higher priority will be given to those who have referred to more manuals related to problem information and peripheral information. [Point 5] We will investigate not only the person seeking advice but all the people who carry out the work to find out about their experience with problem information and related information, and their experience with referring to manuals. This will allow us to match people from the perspective of not only the person seeking advice, but also the person seeking advice. [Point 6] As a result of the pairing process, the system notifies potential consultants who have been given a high priority that they have a problem that the system may be able to help with based on their own experience. [Point 7] In relation to [Point 1] above, in addition to the text information handled by the user terminal UT, the screen images displayed and the paths of the documents used are acquired, and are recorded and stored in association with the text information. When a potential consultant accepts a consultation or receives a notification related to [Point 6], text information including the problem information and peripheral information, which are the pairing conditions, is identified from the stored text information, and the associated screen images and paths are displayed.

[0028] Note that manuals are documents that include a comprehensive full set of information, such as manuals, general procedures that cover all tasks, public notices, etc. In contrast, documents specific to each project, explanatory materials including screen captures, development-related documents, research materials, etc. are referred to as project-specific documents in this embodiment.

[0029] FIG. 2 shows an overview of the process of the search server 1 according to the first embodiment. FIG. 3 shows an overview of the process up to the extraction of the consultation candidates with recommendation levels shown in FIG. 2. A consultant (consultation source user) USX has encountered a problem and is worried about who to consult with. It is assumed that the text information handled by the user terminal UTX used by the consultant USX is stored in advance as work experience (log) in an arbitrary storage unit. When a problem occurs as shown in FIG. 3, the consultant USX issues a search request to the search server 1 via the user terminal UTX, and the search server 1 acquires the text information handled by the user terminal UTX at that time (or a certain time frame including that time) as the current work situation BC. The search server 1 compares the words or words (collectively referred to here as "vocabulary") included in the acquired text information with the accumulated work experience and calculates the frequency of occurrence in the accumulated work experience. The vocabulary distribution image VDX indicates the frequency of occurrence of each word in the text information acquired as the current work situation in the work experience of the consultant USX. Here, light-colored areas indicate a low occurrence frequency, and dark-colored areas indicate a high occurrence frequency.

[0030] Based on the frequency of occurrence, the search server 1 extracts words that the client USX has little experience with from among the words contained in the current business situation of the client USX as "problem information TW," and extracts words that the client has a lot of experience with as "peripheral information PW." The degree of experience can be determined using a pre-specified threshold value.

[0031] Next, the search server 1 judges the amount of experience of the problem information TW and the peripheral information PW for all the business practitioners other than the client. It is assumed that text information handled by each user terminal UT is acquired as the business experience for these business practitioners in advance and stored in an arbitrary storage unit. Specifically, as shown in FIG. 2, the search server 1 extracts consultation candidates who have a lot of experience using the problem information TW, that is, who have a high degree of experience with the problem information TW, from all the business practitioners other than the client. Then, the search server 1 assigns a high recommendation degree to the extracted consultation candidates who have a lot of experience with the peripheral information PW, that is, who have a high degree of experience with the peripheral information PW, and selects the consultation candidates with recommendation degrees. In the example of FIG. 3, users US1 and US2 having vocabulary distribution images VD1 and VD2 are selected as the consultation candidates with recommendation degrees. It is determined that the user US1 has sufficient experience, especially with the peripheral information PW, and the user US2 has little experience with the peripheral information PW but has more experience with the problem information TW.

[0032] Next, as shown in Figure 2, the search server 1 narrows down the prioritized consultation candidates by assigning a high priority to business practitioners who frequently refer to related manuals, i.e., those who have a high degree of experience referring to manuals, from among the selected consultation candidates with recommendation levels.

[0033] If the consultant pairing technology is used to narrow down the "consultation candidates with recommendation levels" by ranking the consultation candidates with those who have a high level of experience with the problem information TW of the consultant USX and those who have a high level of experience with the peripheral information PW at the top, the consultant pairing technology will simply select a consultant who handles a lot of vocabulary in the peripheral information PW. However, the consultant does not simply want to select a consultant who has performed a series of tasks related to the consultant's problem many times, but rather wants to select a consultant who has general knowledge and is likely to know how to deal with the problem even if the consultant does not respond himself. Therefore, in this embodiment, a consultant who knows the work variations and has general knowledge, that is, a consultant who has seen or knows a document "created to be seen by a wide range of people and be useful" and has seen manuals, is selected as a prioritized consultation candidate.

[0034] In the following explanation, for convenience, among all business practitioners (all users) using the search system 100, the consulter (consultation source user) is represented as USX to distinguish it from business practitioners other than the consulter. Similarly, among all user terminals UT, the user terminal used by the consultation source user USX is represented as UTX. However, as should be understood, the consultation source user USX and the user terminal UTX do not refer to any fixed user or user terminal, but may be replaced by other users or user terminals at any time depending on the occurrence of the problem. Hereinafter, when simply referring to the user terminal UT, the user terminal UTX may be included.

[0035] (1-2) Search server (1-2-1) Functional configuration Next, the configuration of the search server 1 according to the first embodiment will be described. Fig. 4 shows an example of a functional configuration of the search server 1. In Fig. 4, the search server 1 has a terminal acquisition content grasping function unit F1 that grasps the handling content of the user terminal UT, a pairing condition generation function unit F2 that generates pairing conditions, a pairing processing function unit F3 that performs pairing processing, and a pairing result output function unit F4 that outputs the pairing result.

[0036] The search server 1 includes a terminal handling content acquisition unit 11 and a terminal handling content storage unit 15 for the terminal acquisition content understanding function unit F1. The search server 1 also includes a pairing condition generation unit 21, a pairing condition editing unit 22, and a pairing condition storage unit 25 for the pairing condition generation function unit F2. The search server 1 also includes a pairing process execution unit 31 and a manual type judgment variable group space storage unit 35 for the pairing process function unit F3. The search server 1 also includes a pairing result presentation unit 41, a pairing result notification unit 42, and a pairing result storage unit 45 for the pairing result output function unit F4.

[0037] The terminal handling content acquisition unit 11 performs a process of acquiring text information handled by each user terminal UT including the user terminal UTX. The terminal handling content acquisition unit 11 can acquire text information displayed on the screen of the user terminal UT by using, for example, UI Automation (hereinafter, referred to as "UIA"), Microsoft Active Accessibility (hereinafter, referred to as "MSAA"), or an interface provided independently by a program running in the user terminal UT. Alternatively, the terminal handling content acquisition unit 11 may acquire information by converting an image displayed on the screen of each user terminal UT into text using character recognition technology. In addition, when a document is displayed on the screen of the user terminal UT, the terminal handling content acquisition unit 11 may acquire text information from metadata of the document. The document may include not only case-specific documents such as emails, web pages, and office documents, but also manuals. Furthermore, the terminal handling content acquisition unit 11 may acquire information by converting a voice signal input / output to / from each user terminal UT into text using voice recognition technology or the like. The terminal handling content acquisition unit 11 acquires text information by converting electronic data into text by applying existing technology to various data formats. The text information acquired by the terminal handling contents acquisition unit 11 is stored in the terminal handling contents storage unit 15, or is passed to the pairing condition generation unit 21 for linkage. Optionally, the terminal handling contents acquisition unit 11 may acquire, together with the text information, a screen image of the entire screen or window from which the text information was acquired, or a path to a document. The terminal handling contents acquisition unit 11 can acquire handling contents not only from a user terminal UTX used by the client USX, but also from a user terminal UT used by a user other than the client.

[0038] The terminal handling content storage unit 15 has a function of storing text information handled by the user terminal UT acquired by the terminal handling content acquisition unit 11, distinguishing between business operators. The terminal handling content storage unit 15 may store the acquired text information in its original form, or may store the text information in a processed form, such as the appearance frequency distribution of words included in the text information. As an option, the terminal handling content storage unit 15 may also store the entire screen or screen image of the window from which the text information was acquired, the document path, etc.

[0039] In response to receiving the trouble occurrence information transmitted from the user terminal UTX in response to the trouble occurrence report operation by the client, the pairing condition generating unit 21 performs a process of comparing the text information handled by the user terminal UTX at the time of the trouble occurrence, acquired by the terminal handling content acquiring unit 11, with the text information stored in the terminal handling content storing unit 15, and extracting words corresponding to the trouble information TW and the peripheral information PW. More specifically, the pairing condition generating unit 21 divides the text information into words, and for each divided word, checks the appearance status in the text information stored in the terminal handling content storing unit 15 of the client himself or other business practitioners, and judges whether it corresponds to the "trouble information TW" or the "peripheral information PW". The pairing condition generating unit 21 stores the judgment result in the pairing condition holding unit 25.

[0040] The pairing condition storage unit 25 stores the vocabulary of the problem information TW and the vocabulary of the peripheral information PW. Optionally, the pairing condition storage unit 25 may also store text information acquired from the user terminal UTX when a problem occurs, as well as a screen image and a document path acquired together with the text information.

[0041] The pairing condition editing unit 22 performs a process of displaying the vocabulary of each of the problem information TW and the peripheral information PW stored in the pairing condition storage unit 25 to the client. For example, the pairing condition editing unit 22 generates information for displaying the information stored in the pairing condition storage unit 25 on a display unit (such as an integrated or external display) of the user terminal UTX, and transmits the information to the user terminal UTX. The pairing condition editing unit 22 also performs a process of deleting, changing, or adding the vocabulary of each of the problem information TW and the peripheral information PW in response to the client's operation inputted via the user terminal UTX. The pairing condition editing unit 22 is optional and may be omitted.

[0042] The manual-type judgment variable group space holding unit 35 holds manual-type judgment variable group space information acquired in advance from actual manuals. This manual-type judgment variable group space information is used when determining whether the document to be acquired for the text information corresponding to the problem information TW and the peripheral information PW that caused the extraction of each consultation candidate is a manual or not, and the details thereof will be described later.

[0043] The pairing process execution unit 31 extracts candidates for consultation users from among users other than the consultation source user according to their work experience, based on the problem information TW and peripheral information PW stored in the pairing condition storage unit 25. More specifically, the pairing process execution unit 31 calculates the following evaluation indexes A and B by comparing the problem information TW and peripheral information PW with text information stored in the terminal handling content storage unit 15 in association with each business practitioner, and extracts people who are determined to have work experience related to the problem based on the evaluation index A as consultation candidates, and then performs processing to assign a higher recommendation level to people with higher evaluation indexes B and A as consultation candidates with a recommendation level. Evaluation index A: Amount of work experience possessed by each person who performs the work on problem information TW Evaluation index B: Amount of work experience possessed by each business practitioner regarding peripheral information PW Furthermore, the pairing process execution unit 31 judges whether the documents to be acquired for the text information corresponding to the problem information TW and the peripheral information PW that caused the extraction of each consultation candidate with recommendation level are manuals or not, based on the manuals judgment variable group space information stored in the manuals judgment variable group space storage unit 35. Then, the pairing process execution unit 31 counts the number of documents that are judged to be manuals, and obtains the number of manuals references. Based on this number of manuals references, the search server 1 becomes able to select prioritized consultation candidates by assigning a higher priority to those who have referred to manuals a lot.

[0044] The pairing result storage unit 45 has a function of storing the prioritized consultation candidates extracted by the pairing process execution unit 31, their respective recommendation levels, and the number of times manuals were referenced. As an option, the pairing result storage unit 45 can also store, for each prioritized consultation candidate, the degree of experience of the candidate with the problems and the business situation at the time of the problem occurrence, such as the frequency of occurrence in the accumulated text information for each word in the vocabulary of the problem information TW and the vocabulary of the peripheral information PW.

[0045] The pairing result presentation unit 41 performs a process of displaying the data stored in the pairing result storage unit 45 to the client. For example, the pairing result presentation unit 41 generates information for displaying the pairing result on a display unit (such as an integrated or external display) of the user terminal UTX of the client user USX, and transmits the information to the user terminal UTX. As an option, the pairing result presentation unit 41 may generate display information for starting communication (telephone, e-mail, business chat, etc.) with a prioritized consultation candidate in response to a selection operation by the client, and transmit the information to the user terminal UTX.

[0046] The pairing result notification unit 42 performs a process of notifying the occurrence of a problem to the business practitioners whose rankings, when arranged in order of the degree of experience of referring to manuals, are within a pre-specified threshold value among the prioritized consultation candidates extracted as a result of the pairing process executed by the pairing process execution unit 31. For example, the pairing result notification unit 42 determines the business practitioners whose rankings are within the above threshold value as consultation users who are final consultation candidates, and notifies the user terminals UT of the consultation users that they have been extracted as candidates, the occurrence of a problem, or information about the consultation source user. As an option, the pairing result notification unit 42 may further notify the text information acquired by the terminal UTX of the client when the problem occurs and stored in the pairing condition storage unit 25, as well as information for displaying the screen image acquired together with the text information and the path of the document. As an option, the pairing result notification unit 42 may also identify the text information of the final consultation candidate stored in the terminal handling content storage unit 15 that includes the problem information TW and the peripheral information PW, and notify the screen image and the path associated with it. As a further option, the pairing result notification unit 42 may generate display information that initiates communication (telephone call, email, business chat, etc.) with the client in response to a consultation reception operation by the final consultation candidate, and transmit the display information to the user terminal UT.

[0047] (1-2-2) Hardware Configuration Fig. 5 shows a schematic example of the hardware configuration of the search server 1. As shown in Fig. 5, the search server 1 includes a processor 101, a RAM 102, a program memory 103, a storage device 104, and an input / output interface 105, which are connected to each other via a bus 106. The processor 101 controls the RAM 102, the program memory 103, the storage device 104, and the input / output interface 105, and exchanges signals with them.

[0048] The processor 101 includes a general-purpose circuit such as a CPU. The RAM 102 is used by the processor 101 as a working memory. The RAM 102 includes a volatile memory such as an SDRAM. The program memory 103 stores programs executed by the processor 101, including an information generating program. The programs include computer-executable instructions. As the program memory 103, for example, a ROM is used.

[0049] The processor 101 loads the program stored in the program memory 103 into the RAM 102, and interprets and executes the program. When the information generation program is executed by the processor 101, it can cause the processor 101 to execute the processing functions of the terminal handling details acquisition unit 11, the pairing condition generation unit 21, the pairing condition editing unit 22, the pairing process execution unit 31, the pairing result presentation unit 41, and the pairing result notification unit 42, which are described above with respect to the search server 1.

[0050] The storage device 104 can function as a storage area for the above-mentioned terminal handling content accumulation unit 15, pairing condition holding unit 25, and pairing result holding unit 45. The storage device 104 includes a non-volatile memory such as a hard disk drive (HDD) or a solid state drive (SSD). A part of the area of ​​the storage device 104 may be used as the program memory 103.

[0051] The input / output interface 105 includes a communication module for communicating with an external device and a plurality of terminals for connecting peripheral devices. The external device is, for example, a user terminal UT. The communication module may include, for example, a wired module and / or a wireless module. Examples of peripheral devices include a display, a keyboard, and a mouse.

[0052] The processor 101 may include multiple processors. At least a part of the processing functions may be realized in various other formats including integrated circuits such as an ASIC (Application Specific Integrated Circuit) and an FPGA (Field Programmable Gate Array). At least a part of the memory area of ​​the storage device 104 may be provided in a device different from the search server 1.

[0053] (2) Operation (2-1) Advance preparation The following explanatory variable(s) are used as indicators for determining whether a document is a manual or a case-specific document. Explanatory variable group V1: Conformance rate with manual format Explanatory variable V2: Range of file users Explanatory variables V3: Period of use in the file, frequency of use, etc. Here, the explanatory variable group V1 is a value obtained by counting the number of occurrences of words such as "update history," "version number," and "modifier" that frequently appear in manuals and the like. The explanatory variable V2 indicates, as a number, how many organizations the file is used in. The larger the value, the more likely it is that the document is used by multiple organizations and is therefore a manual. The explanatory variable group V3 indicates the period of use, the number of times of use, the frequency of use, and periodicity as numerical values. For example, the period of use is the number of days or years from when the use started to when, and the number of times of use is the number of times it has been viewed. The frequency of use is the number of times it has been viewed in a period such as one month, and the periodicity can be the peak number of times of use per month or every three months (Q).

[0054] For documents that are to be classified as manuals or case-specific documents, these three explanatory variables (groups), i.e., explanatory variable group V1, explanatory variable group V2, and explanatory variable group V3, are considered together. Decision variable group = (explanatory variable group V1, explanatory variable V2, explanatory variable group V3) This group of decision variables is used to determine whether a document is a manual or a case-specific document.

[0055] As a preliminary step for this, real data is read in to grasp in advance what kind of distribution (space) the judgment variable group of actual manuals will be, and what kind of distribution (space) the judgment variable group of actual case-specific documents will be. That is, a large number of actual manuals and case-specific documents are prepared, and the distribution of the judgment variable group in the manuals and the judgment variable group in the case-specific documents are calculated and statistically analyzed. Then, the manuals and the case-specific documents are clustered, respectively. Hereinafter, the judgment variable group space of manuals is denoted as α, and the judgment variable group space of case-specific documents is denoted as β. Information on the clustered manual judgment variable group space α and case-specific document judgment variable group space β is stored in the manual judgment variable group space storage unit 35.

[0056] (2-2) Overview of the overall operation Next, an information processing operation by the search server 1 serving as the information processing device according to the first embodiment configured as above will be described.

[0057] FIG. 6 is a diagram showing an outline of the information processing operation. In step S1, the search server 1 performs a process of grasping the terminal handling contents. The process of grasping the terminal handling contents includes acquiring text information (first text information) handled in the user terminal UTX in response to a search request issued by the consultant (first user) when a problem occurs. The process of grasping the terminal handling contents may further include acquiring text information (second text information) as the business experience of the consultant in the consulting user terminal UTX, and acquiring text information (third text information) as the business experience of all user terminals UT except the consulting user terminal UTX.

[0058] In step S2, the search server 1 monitors the occurrence of a problem. If it is determined in step S2 that no problem has occurred, the process proceeds to step S1, and if it is determined that a problem has occurred, the process proceeds to step S3. For example, the search server 1 executes subsequent processing in response to a search request issued from any of the user terminals UT to search for someone to consult with. The search request is sent from the user terminal UTX to the search server 1, for example, when the user who wishes to consult presses a "consult button" displayed as a GUI on the display of the user terminal UTX.

[0059] In step S3, the search server 1 performs a process of generating pairing conditions. The process of generating pairing conditions includes a process of dividing words included in the acquired first text information into a first group of words whose frequency of occurrence is lower than a predetermined threshold and a second group of words whose frequency of occurrence is equal to or higher than a predetermined threshold according to the frequency of occurrence in the second text information. The first group of words corresponds to "problem information TW" and includes words that the consulter has less experience with. The second group of words corresponds to "peripheral information PW". Note that the "frequency of occurrence" here is calculated as the number of times a certain word appears in the text information to be compared divided by the total number of all words included in the text information to be compared. In the example of step S3, the frequency of occurrence is calculated by calculating the number of times each word included in the acquired first text information appears in the second text information and dividing the number by the total number of all words included in the second text information.

[0060] In step S4, the search server 1 performs a pairing process. The pairing process includes a process of extracting consultation destination user candidates from users other than the consultation source user by regarding the appearance frequency of the problem information TW and the peripheral information PW as the degree of work experience, and calculating the recommendation level of each. The pairing process also includes a process of determining whether the document in which the problem information TW and the peripheral information PW appear is a manual or the like, and counting the number of references to the manual or the like.

[0061] In step S5, the search server 1 outputs the pairing result. The output of the pairing result includes a process of displaying the pairing result to the consultation source user and the consultation destination user.

[0062] The above step S3 will be described in more detail below. In the following description, · Set of all business operators (including consultants and consultant recipients): K · A set of records for each practitioner k (∈K): L k · A set of vocabulary included in the problem information TW (problem information vocabulary set): T - Set of vocabulary included in the surrounding information PW (surrounding information vocabulary set): C Let us assume that.

[0063] The search server 1 searches for each word t included in the problem information vocabulary set T. i For each word, the frequency of occurrence of the corresponding word is: m i,k Then, calculate the following as the "amount of work experience related to problem information TW".

[0064]

number

[0065] The coverage of the problem information TW, in other words, represents the ratio of words belonging to the first word group (vocabulary of the problem information TW) that exist in the records of each business practitioner. The handling of the problem information TW, in other words, represents the sum of the frequencies at which words belonging to the first word group appear in the records of each business practitioner divided by the total number of words belonging to the first word group.

[0066] The search server 1 also searches for each word c j For each word, the frequency of occurrence of the corresponding word n ​​is calculated from the vocabulary frequency distribution. j,k Then, calculate the following as the "amount of work experience related to peripheral information PW".

[0067]

number

[0068] The coverage of the peripheral information PW, in other words, represents the proportion of words belonging to the second word group (vocabulary of the peripheral information PW) that exist in the records of each business practitioner. The handling of the peripheral information PW, in other words, represents the sum of the frequencies at which words belonging to the second word group appear in the records of each business practitioner divided by the total number of words belonging to the second word group.

[0069] Furthermore, the search server 1 extracts consultation candidates from the set K using the "amount of work experience related to the problem information TW", and for each consultation candidate k, calculates the "recommendation degree r k ” is calculated and assigned to each consulting candidate k.

[0070]

number

[0071] In the above formula, γ, δ, ε, and ζ are parameters that control which of the evaluation indicators related to “amount of work experience related to problem information TW” and “amount of work experience related to peripheral information PW” are reflected in the recommendation level and to what extent, and are specified in advance by, for example, an administrator of the search server 1.

[0072] In other words, the recommendation degree is a weighted sum of the coverage of the problem information TW, the handling of the problem information TW, the coverage of the peripheral information PW, and the handling of the peripheral information PW for each consultation candidate k. The recommendation degree is higher as the coverage of the problem information TW is higher, the handling of the problem information TW is higher, the coverage of the peripheral information PW is higher, and the handling of the peripheral information PW is higher.

[0073] In addition, it is not necessary to use all of the coverage of the problem information TW, the handling of the problem information TW, the coverage of the peripheral information PW, and the handling of the peripheral information PW in calculating the recommendation level, and any one of them may be used for calculation. In other words, the above γ, δ, ε, and ζ may be positive values ​​or zero, and at least one of γ, δ, ε, and ζ may be non-zero. The operation of the search server 1 will be explained in further detail below.

[0074] (2-3) Understanding of terminal handling details FIG. 7 is a diagram showing details of the process in step S1 shown in FIG. In step S11, the terminal handling content acquisition unit 11 acquires text information handled by each user terminal UT. The text information may be acquired only from the screen displayed to the user on the display of the user terminal UT, or may be acquired from an active electronic file on the user terminal including a portion not displayed on the screen. At this time, manuals and case-specific documents are not distinguished as electronic files.

[0075] As described above, the terminal handling content acquisition unit 11 may acquire text information displayed on the screen using UIA, MSAA, or an interface provided independently by a program running in the user terminal UT. Alternatively, the terminal handling content acquisition unit 11 may convert images displayed on each terminal screen into text using character recognition technology, or may convert voice signals input / output to / from each user terminal UT into text using voice recognition technology or the like. The terminal handling content acquisition unit 11 acquires text information by converting data into text by applying existing technology to various data formats. As an option, the terminal handling content acquisition unit 11 may acquire, together with the text information, a screen image of the entire screen or window from which the text information was acquired, or a path to a document.

[0076] Fig. 8 shows an example of text information acquired by the terminal handling content acquisition unit 11. Here, the order content registration process is being performed on the user terminal UT, and screen DS1 is a capture of the display on the display of the user terminal UT at that time. Screen DS1 includes text information such as "order content", "order slip number", etc. Based on this screen display, text information TI1 is acquired as described above.

[0077] 7, the terminal handling content acquisition unit 11 distinguishes the business operator by associating the acquired text information with a user ID or the like, and stores the information in the terminal handling content storage unit 15. The text information may also be associated with date and time information.

[0078] In this way, the terminal handling content accumulation unit 15 accumulates information every time text information is acquired by the terminal handling content acquisition unit 11. Alternatively, as a prerequisite for starting the operation of the search server 1, a certain amount or more of text information regarding all users may be accumulated in advance in the terminal handling content accumulation unit 15, regardless of whether or not a problem has occurred.

[0079] Here, the terminal handling content storage unit 15 may store text information in a processed form, such as the frequency distribution of vocabulary included in the text information. Optionally, the terminal handling content storage unit 15 may store the entire screen or window image of the user terminal UT from which the text information was acquired, or the path of the target document, in association with the text information, together with the acquisition time. The data may be distributed and stored in an area on a data storage device accessible only to each business performer of each user terminal UT, or data for all business performers may be stored together on a file server or the like. Alternatively, the method of storage may be changed depending on the type of data, such as distributing the text information, screen images, and document paths and collectively arranging only the frequency distribution of vocabulary.

[0080] Fig. 9 shows an example of data stored in terminal handling content storage unit 15. The example in Fig. 9 shows an example in which text information is stored for each business implementer (business implementer 1,...,business implementer k) with the timing of its acquisition distinguished. In this example, the items include "acquisition time", "text information", "screen image", and "pass". However, acquisition and storage of the acquisition time, screen image, and pass are optional.

[0081] (2-4) Pairing condition generation process FIG. 10 is a diagram showing details of the process in step S3 shown in FIG. In step S31, the pairing condition generating unit 21 acquires text information acquired by the terminal handling content acquiring unit 11 that is being handled by the user terminal UTX when a problem occurs.

[0082] In step S32, the pairing condition generating unit 21 divides the acquired text information into morphemes by morpheme analysis. Among the divided morphemes, particles and other morphemes with low importance may be removed. Alternatively, the morphemes may be integrated to form more complex words by referring to a dictionary registered in advance. Each piece of information obtained by the division is referred to as a word or phrase here. Note that the morpheme analysis is an example, and division may be performed by other methods.

[0083] In step S33, the pairing condition generating unit 21 checks the appearance status of each of the obtained words in the text information stored in the terminal handling content storing unit 15, and judges whether or not the words correspond to "problem information TW" and "peripheral information PW". Here, the text information to be compared with each word may be text information stored about the caller or other service practitioners. The pairing condition generating unit 21 judges "problem information TW" and "peripheral information PW" using, for example, the following method.

[0084] (2-4-1) First Example of Processing for Determining "Problem Information TW" and "Peripheral Information PW" (2-4-1-1) First Step For each word obtained in step S32, the number of times it appears in the text information of the client himself stored in the terminal handling content storage unit 15 is checked to find the frequency of appearance.

[0085] (2-4-1-2) Second step Next, the occurrence frequency of each word obtained in (2-4-1-1) above is compared with a preset threshold value to determine whether it is high or low. For example, words with an occurrence frequency equal to or higher than the threshold value are determined as "peripheral information PW," and words with an occurrence frequency lower than the threshold value are determined as "problem information TW."

[0086] Fig. 11 is a diagram illustrating a judgment according to the first embodiment. For each word acquired from the client's terminal UTX, a word with a high frequency of occurrence is judged as "peripheral information PW" and a word with a low frequency of occurrence is judged as "problem information TW" according to the frequency of occurrence in the client's accumulated past text information.

[0087] (2-4-2) Second Example of Processing for Determining "Problem Information TW" and "Peripheral Information PW" (2-4-2-1) First step For each word obtained in step S32, the number of times it appears in the text information of the client himself stored in the terminal handling content storage unit 15 is checked to find the frequency of appearance.

[0088] (2-4-2-2) Second step The frequency of occurrence of each word of the client obtained in (2-4-2-1) above is compared with a threshold value to determine whether it is high or low.

[0089] (2-4-2-3) Third step For each word obtained in step S32, the number of times it appears in the text information stored in the terminal handling content storage unit 15 for each service provider other than the consultant is checked to determine the frequency of appearance.

[0090] (2-4-2-4) Fourth step The frequency of occurrence of each word for each practitioner obtained in (2-4-2-3) above is compared with a threshold value to determine whether it is high or low.

[0091] (2-4-2-5) Fifth step Count the number of practitioners judged as "high" and the number of practitioners judged as "low" obtained in (2-4-2-4) above.

[0092] (2-4-2-6) Sixth Step Based on the results of (2-4-2-2) and (2-4-2-5) above, it is determined whether each word corresponds to "problem information TW" and "peripheral information PW". Note that the threshold values ​​used in (2-4-2-2) and (2-4-2-4) may be the same or different.

[0093] FIG. 12 is a diagram illustrating a judgment according to the second embodiment. As with the first embodiment, for each word acquired from the client's terminal UTX, a word with a high frequency of occurrence is judged as "peripheral information PW" and a word with a low frequency of occurrence is judged as "problem information TW" according to the frequency of occurrence in the client's accumulated past text information. However, unlike the first embodiment, a word with a low frequency of occurrence in the text information of other business practitioners is excluded from the "peripheral information PW" and "problem information TW". In this way, by excluding words with a low frequency of occurrence for all other business practitioners, it is possible to exclude words (e.g., IDs of new cases) that are unique only when a problem occurs and are not useful for searching for past response cases or response experiences throughout the entire past work of all business practitioners.

[0094] (2-4-3) Example of determining threshold The threshold values ​​used in the above (2-4-1-2), (2-4-2-2) and (2-4-2-4) can be determined, for example, as follows. [Example 1] A person such as the administrator of search server 1 specifies this as a fixed value in advance. [Example 2] Calculate the statistical values ​​(average, median, mode, variance, standard deviation, etc.) of the frequency of occurrence of vocabulary in the client's accumulated text information, and make a decision from those statistical values. [Example 3] Calculate the statistical values ​​(average, median, mode, variance, standard deviation, etc.) of the frequency of occurrence of vocabulary in the accumulated text information of all business practitioners, and make a determination from those statistical values.

[0095] (2-4-4) Example of saving judgment results Next, in step S34 of FIG. 10, the pairing condition generating unit 21 stores the determination result in the pairing condition holding unit 25.

[0096] The pairing condition storage unit 25 stores the vocabulary of the problem information TW and the vocabulary of the peripheral information PW. Optionally, the pairing condition storage unit 25 may store text information acquired when the problem occurs, as well as a screen image and a document path acquired together with the text information.

[0097] 13 shows an example of data stored in the pairing condition storage unit 25. In this example, the items include "pairing process ID," "consultant," "problem information TW," "peripheral information PW," "text information," "screen image," and "pass." Note that "text information," "screen image," and "pass" are information of the same names as the data stored in the terminal handling content storage unit 15, and all of them are optional.

[0098] As an option, the information stored in the pairing condition storage unit 25 can be edited by the client. As described above, the pairing condition editing unit 22 enables such editing. For example, every time the information stored in the pairing condition storage unit 25 is updated or at regular intervals, the pairing condition editing unit 22 generates and transmits information for displaying each word of the problem information TW and the peripheral information PW stored in the pairing condition storage unit 25 to the user terminal UTX used by the client. When the client performs an operation to request deletion, change, addition, or the like of a word via the user terminal UTX, the request is transmitted to the search server 1, and the pairing condition editing unit 22 edits the information stored in the pairing condition storage unit 25 in response to the client's request.

[0099] (2-5) Pairing process 14 is a diagram showing details of the process of step S4 shown in FIG. 6. This process is mainly executed by the pairing process execution unit 31. As described above, the pairing process execution unit 31 compares the problem information TW and peripheral information PW stored in the pairing condition storage unit 25 with text information stored in association with each service provider, Evaluation index A: Amount of work experience possessed by each person who performs the work on problem information TW Evaluation index B: Amount of work experience possessed by each business practitioner regarding peripheral information PW The pairing process execution unit 31 then extracts people who are determined to have work experience related to the problem based on the evaluation index A, and assigns a higher recommendation level to people with larger evaluation indexes B and A. Furthermore, the pairing process execution unit 31 counts the number of times related manuals have been referenced for each person to whom a high recommendation level has been assigned.

[0100] First, in step S41, the pairing process execution unit 31 reads out the problem information TW of the client and the peripheral information PW from the pairing condition storage unit 25.

[0101] In step S42, the pairing process executing unit 31 extracts, based on the evaluation index A, business practitioners who have a lot of experience using the problem information TW as consultation candidates.

[0102] In step S43, the pairing process executing unit 31 assigns a high recommendation level to a service practitioner who knows the problem information TW and has a lot of experience with the peripheral information PW.

[0103] In step S44, the pairing process execution unit 31 counts the number of times that a service provider who knows the problem information TW and has a lot of experience with the peripheral information PW, that is, a service provider to whom a high recommendation level is assigned, has referred to the manual.

[0104] In step S45, the pairing process execution unit 31 stores the pairing result in the pairing result storage unit 45.

[0105] The pairing process is described in more detail below. As mentioned above, in the following description: Set of practitioners: K A set of vocabulary included in the problem information TW (problem information vocabulary set): T A set of vocabulary words included in the surrounding information PW (surrounding information vocabulary set): C is used.

[0106] (2-5-1) First Example of Pairing Process: Example of Processing When Text Information is Stored with Different Acquisition Timings The text information of each business operator stored in the terminal handling content storage unit 15 is held as a different “record” (row in FIG. 9) due to the difference in acquisition timing. Let L be the set of records for each business operator k (∈K). k Let us assume that.

[0107] In this first embodiment, the pairing process execution unit 31 checks to what extent both the problem information TW and the peripheral information PW are included for each record, not for the entire accumulated text information of each business practitioner. This makes it possible to consider the extent to which the business practitioner has experience of handling the problem information TW and the peripheral information PW at the same time, that is, the experience of the same business situation as the client.

[0108] If a timestamp is assigned to each record, the timing at which the text information was acquired may be taken into consideration, for example, by limiting the records to those within a certain period from when the problem occurred.

[0109] (2-5-1-1) First Step The pairing process execution unit 31 performs pairing for each record l (∈L) of the text information of each business practitioner k (∈K) stored in the terminal handling content storage unit 15. k ) the following (i) and (ii) shall be carried out.

[0110] (i) The pairing process execution unit 31 performs pairing for each word t included in the problem information vocabulary set T. i For each word, the frequency of occurrence m i,k,l Then, calculate the following:

[0111]

number

[0112] (ii) The pairing process execution unit 31 performs pairing for each word c included in the peripheral information vocabulary set C. j For each word, the frequency of occurrence of n j,k,l Then, calculate the following:

[0113]

number

[0114] (2-5-1-2) Second step The pairing process execution unit 31 judges whether each record l of the text information of each business practitioner k is a record related to a problem by the coverage degree g of the problem information TW in the record. k,l and its threshold g * are compared and judged as follows:

[0115]

number

[0116] (2-5-1-3) Third step The pairing process execution unit 31 performs the following (iii) and (iv) for each business practitioner k.

[0117] (iii) Calculate the amount of work experience related to problem information TW as follows:

[0118]

number

[0119] (iv) Calculate the amount of work experience related to peripheral information PW as follows:

[0120]

number

[0121] (2-5-1-4) Fourth step The pairing process execution unit 31 determines whether each business practitioner k has business experience related to the problem.

[0122]

number

[0123] Using the "amount of work experience related to problem information TW," the judgement is made based on whether or not the person is judged to have work experience related to problems in either or both of the following.

[0124]

number

[0125] (2-5-1-5) Fifth step The pairing process execution unit 31 calculates a "recommendation degree r" for each business practitioner k who is determined to have business experience related to the problem in the fourth step (2-5-1-4) above. k " is calculated.

[0126] (2-5-1-5-1) First Example of Recommendation Level Calculation Method The pairing process execution unit 31 calculates a "recommendation degree r" for each business practitioner k using the "amount of business experience related to the problem information TW" and the "amount of business experience related to the peripheral information PW" as follows: k " can be calculated.

[0127]

number

[0128] Here, γ, δ, ε, and ζ are parameters that control which of the evaluation indexes related to "amount of work experience related to problem information TW" and "amount of work experience related to peripheral information PW" are reflected in the recommendation level and to what extent, as described above, they are specified in advance by, for example, an administrator of the search server 1. γ, δ, ε, and ζ can take zero or positive values, and at least one of them takes a non-zero value.

[0129] (2-5-1-5-2) Second Example of Recommendation Level Calculation Method The pairing process execution unit 31 performs pairing for each word t included in the problem information vocabulary set T. i and each word c in the surrounding information vocabulary set C j The following weights are calculated for each business practitioner k. This corresponds to the IDF (Inverse Document Frequency) when each record l of the accumulated text information of each business practitioner k is regarded as one document.

[0130]

number

[0131] The pairing process execution unit 31 further calculates, as the "weighted amount of business experience related to the problem information TW",

[0132]

number

[0133] Calculate the weighted amount of work experience related to peripheral information PW.

[0134]

number

[0135] For each practitioner k who is determined to have work experience related to the problem in the fourth step above, the "recommendation degree r k " is calculated as follows:

[0136]

number

[0137] (2-5-2) Counting the number of times manuals were referenced The pairing process execution unit 31 counts the number of times manuals are referenced for a consultation candidate to which a high recommendation degree is assigned among the consultation candidates with recommendation degrees extracted / calculated as described above. Fig. 15 is a flowchart showing an example of a procedure for counting the number of times manuals are referenced by a business practitioner in step S44 of Fig. 14.

[0138] (2-5-2-1) Select the person to be processed In step S441, the pairing process executing unit 31 selects one business practitioner to which a high recommendation level is assigned from among the consultation candidates with recommendation levels. For example, the pairing process executing unit 31 can rank the business practitioners with recommendation levels according to their recommendation levels, and determine the consultation candidates with recommendation levels up to a predetermined rank as the business practitioners to which a high recommendation level is assigned. Alternatively, the pairing process executing unit 31 can determine the consultation candidates with recommendation levels whose recommendation level values ​​are equal to or greater than a predetermined value as the business practitioners to which a high recommendation level is assigned. The predetermined rank and the predetermined value are designated in advance, for example, by an administrator of the search server 1, etc.

[0139] (2-5-2-2) Documents to be processed In step S442, the pairing process execution unit 31 selects one document that the selected processing target has referenced within a predetermined period from the terminal handling content storage unit 15. This selection can be made by determining whether or not the "acquisition time" in the "record" (line in FIG. 8) that holds the text information of the business executor who is the selected processing target, which is stored in the terminal handling content storage unit 15, is within a predetermined period. The predetermined period is designated in advance by, for example, an administrator of the search server 1.

[0140] (2-5-2-3) Manuals Judgment In step S443, the pairing process execution unit 31 determines whether the selected document is a manual. In this determination, an explanatory variable (group) is calculated, and a judgment variable group is calculated from the calculated explanatory variable (group). Then, whether the calculated judgment variable group is within the judgment variable group space α of manuals that has been clustered in advance is used to determine whether the document is a manual. If the pairing process execution unit 31 determines that the document is a manual, it proceeds to step S444, and if it determines that the document is not a manual, it proceeds to step S445.

[0141] (2-5-2-3-1) First step The pairing process execution unit 31 calculates an explanatory variable set V1 (a conformance rate with the manual format). For example, the pairing process execution unit 31 can calculate the explanatory variable set V1 by counting the occurrences of predetermined words such as "update history," "version number," and "modifier" stored in the record for the selected processing target document in the terminal handling content storage unit 15.

[0142] (2-5-2-3-2) Second step The pairing process execution unit 31 calculates the explanatory variable V2 (range of file users). For example, the pairing process execution unit 31 can calculate this explanatory variable V2 by extracting business practitioners who have used the selected document to be processed, determining the organizations to which the extracted business practitioners belong, and counting the number of organizations. Whether or not each business practitioner has used the document to be processed can be determined based on whether or not the file of the document to be processed is recorded in the "path" in the data for each business practitioner in the terminal handling content accumulation unit 15. In addition, the organization to which each business practitioner belongs can be determined by referring to the business practitioner's affiliation information accumulated in a user database (not shown).

[0143] (2-5-2-3-3) Third step The pairing process execution unit 31 calculates the explanatory variable set V3 (the usage period, usage frequency, etc., of the file). For example, the pairing process execution unit 31 calculates when the file of the selected document to be processed is used, how many times it has been used, for example, how many times per month it is used, and how many times the usage frequency peaks every three months (Q) as a periodicity. These can also be counted by referring to the data for each business practitioner in the terminal handling content accumulation unit 15 for the document to be processed.

[0144] (2-5-2-3-4) Fourth step The pairing process execution unit 31 considers the explanatory variable group V1, explanatory variable V2, and explanatory variable group V3 calculated in the first to third steps together and determines the pairing process result for the document to be processed. Decision variable group = (explanatory variable group V1, explanatory variable V2, explanatory variable group V3) Calculate.

[0145] (2-5-2-3-5) Fifth step The pairing process execution unit 31 calculates whether the judgment variable group of the document to be processed calculated in the fourth step is in the judgment variable group space α of manuals, in the judgment variable group space β of case-specific documents, or which space is closer in terms of distance even if it is not in either space. There is no particular limitation on the calculation method. The pairing process execution unit 31 determines that the document to be processed is a document in the space or in the space that is closer in distance. As described above, information on the judgment variable group space α of manuals and the judgment variable group space β of case-specific documents is stored in advance in the manual judgment variable group space storage unit 35.

[0146] (2-5-2-4) Counting the word groups that make up the business situation In step S444, the pairing process execution unit 31 counts how many of the target documents determined to be manuals contained "word groups constituting the business situation" included in the screen the caller was referring to when the problem occurred. The pairing process execution unit 31 adds up the number of "word groups constituting the business situation" included in the counted target documents, that is, the number of "word groups included in the problem information TW" and "word groups included in the peripheral information PW," and determines this as the number of manuals referenced.

[0147] (2-5-2-5) Update document to be processed In step S445, the pairing process execution unit 31 judges whether or not all documents referenced within a predetermined period by the business practitioner selected as the processing target in step S441 in the terminal handling content storage unit 15 have been selected as processing targets. If there are documents that have not yet been selected as processing targets, the pairing process execution unit 31 proceeds to step S442, selects the unselected documents, and repeats the above steps. On the other hand, if all documents have been selected as processing targets, the pairing process execution unit 31 proceeds to step S446.

[0148] (2-5-2-6) Update of processing target In step S446, the pairing process execution unit 31 judges whether or not all of the business practitioners to which a high recommendation level has been assigned have been selected as processing targets. If there is a business practitioner who has not yet been selected as a processing target, the pairing process execution unit 31 proceeds to step S441, selects the unselected business practitioner, and repeats the above steps. If all business practitioners have been selected as processing targets, the pairing process execution unit 31 ends the procedure of counting the number of times the business practitioners have referred to manuals in step S44.

[0149] (2-5-3) Saving pairing results The business practitioners of the prioritized consultation candidates extracted / calculated by the pairing process execution unit 31 as described above, and their number of times of reference to manuals and their recommendation levels are stored in the pairing result storage unit 45 as pairing results in step S45 of FIG. 14.

[0150] FIG. 16 shows an example of information stored in the pairing result storage unit 45. As shown in FIG. 16, the pairing result storage unit 45 stores "consultation candidates" and their respective "number of times manuals are referenced" and "recommendation levels" in association with a "pairing process ID" that identifies each pairing process. In the example of FIG. 16, the "consultation candidates" are stored as user IDs, but they may also be stored by name, nickname, or the like.

[0151] As an option, the pairing result storage unit 45 can also store "problem information TW" and "peripheral information PW" for each consultation candidate, as shown in Fig. 16. In the example of Fig. 16, for the record of "pairing process ID = PR0001", "consultation candidate = User0051" is calculated as "manual reference count = 15" and "recommendation level = 23.2", and the "amount of business experience" related to the problem information TW is "coverage = 1.0" and "handling = 3.8", and the occurrence frequency by vocabulary is "monthly deployment = 126" and "total performance at the end of this month = 103", and similarly, the "amount of business experience" related to the peripheral information PW is "coverage = 1.0" and "handling = 8.2", and the occurrence frequency by vocabulary is "SMART = 418", "case ID = 356", and "order case ID = 301". As an option, the pairing result storage unit 45 may also store, for each of the "problem information TW" and "peripheral information PW," the frequency of appearance in the accumulated text information, the candidate's degree of experience with the problem and the work situation at the time the problem occurred, etc.

[0152] (2-6) Pairing result output process FIG. 17 is a diagram showing details of the process in step S5 shown in FIG.

[0153] In step S51, the pairing result presentation unit 41 and the pairing result notification unit 42 each read out the pairing result stored in the pairing result storage unit 45.

[0154] In step S52, the pairing result presenting unit 41 generates and outputs presentation information for presenting the pairing result to the client. The output presentation information is transmitted to the user terminal UTX used by the client via the network NW, for example. When the user terminal UTX receives the presentation information, it generates a corresponding display screen, displays it on a display or the like, and presents it to the client. The presentation information may include information for displaying contact information of the prioritized consultation candidates on the display screen. As an option, when the client selects the prioritized consultation candidates via the display screen of the user terminal UTX, communication (telephone, e-mail, business chat, etc.) with the selected prioritized consultation candidates may be started from the user terminal UTX of the client. The communication between the client and the prioritized consultation candidates may be via the search server 1 or may not be via the search server 1.

[0155] In step S53, the pairing result notification unit 42 ranks the prioritized consultation candidates based on the priority, and generates and outputs notification information for notifying the occurrence of a problem to the "final consultation candidates" who are business implementers whose ranks in the order of priority are within a pre-specified threshold. The output notification information is transmitted, for example, via the network NW to the user terminal UTY, which is the user terminal UT used by each final consultation candidate. When the user terminal UTY receives this notification information, it generates a corresponding display screen, displays it on a display, etc., and presents it to each final consultation candidate. As an option, the output notification information may include information for displaying text information acquired by the user terminal UTX of the client when a problem occurs and stored in the pairing condition storage unit 25, as well as screen images acquired together with the text information, document paths, etc. Also as an option, the output notification information may include information for identifying text information of the final consultation candidates stored in the terminal handling content storage unit 15 that includes problem information TW and peripheral information PW, and displaying screen images and paths associated with the text information.

[0156] Furthermore, as an option, when a consultation reception operation by the final consultation candidate is accepted via the display screen of the user terminal UTY, communication (telephone, e-mail, business chat, etc.) with the person seeking advice may be started from the user terminal UTY of the final consultation candidate. Again, the communication between the final consultation candidate and the person seeking advice may be via the search server 1 or may not be via the search server 1.

[0157] It should be noted that the above steps S52 and S53 may be executed at any timing, either simultaneously or in parallel, or sequentially.

[0158] As described above in detail, in the first embodiment of the present invention, in response to a search request issued by a first user (consultation user) USX to search for candidate users from among a plurality of users, first text information handled by the terminal UTX used by the first user is acquired. Then, for words included in the first text information, the occurrence frequency in second text information acquired from a terminal device used in the past by the first user is calculated. The words included in the first text information are classified into a first word group (problem information TW) whose occurrence frequency is lower than a first threshold and a second word group (peripheral information PW) whose occurrence frequency is equal to or higher than the first threshold according to the occurrence frequency in the second text information. Next, for words belonging to the first word group, the occurrence frequency in third text information related to each of the past tasks of the plurality of users to be searched is determined, and candidate users are extracted according to the occurrence frequency. Furthermore, the manuals that each of the extracted candidate users has referred to in the past are determined, and the priority of each candidate user is calculated so that the higher the frequency of occurrence of words belonging to the first word group in the manuals, the higher the priority. Then, information about the candidate users including the priority is presented to the first user who is the consultation source user. The frequency of occurrence of words belonging to the second word group can also be used to extract the candidate users. Furthermore, for the extracted candidate users, a recommendation degree is calculated based on the frequency of occurrence of words in the first word group (problem information TW) and / or the second word group (peripheral information PW), and the candidate users to whom a priority should be given can be narrowed down based on this recommendation degree. The consultation source user may be presented with the candidate users together with the calculated priority or in an order according to the calculated priority. Furthermore, the candidate users may also be notified of information about the consultation source user, the problem information TW, or the peripheral information PW.

[0159] (3) Effects As described above in detail, according to the first embodiment, text information (vocabulary) handled in the work of all business practitioners is accumulated for each business practitioner, and when a certain person in charge encounters a problem during work, for example, by pressing the consultation button on the screen, the vocabulary in the business screen is extracted, the vocabulary of the problem information TW and the peripheral information PW are determined, and a consultation destination that has experience and has frequently referred to related manuals is searched for and presented to the person in charge. More specifically, the determination of the problem information TW and the peripheral information PW is performed by determining, among the vocabulary in the business screen, vocabulary that appears less frequently in the text information accumulated by the person in charge (unfamiliar with little experience in the past) as the problem that the person in charge wants to consult about, and determining the other vocabulary as the peripheral information PW. In addition, the search for consultation partners searches for consultation partners who have a high coverage and frequency of occurrence of problem vocabulary in the accumulated text information (they have a lot of experience with the problem of the person seeking advice) and a high coverage and frequency of occurrence of peripheral information PW vocabulary (they have similar work experience to the person seeking advice), and from among the searched consultation partners, the system recommends consultation partners who have a high number of references to related manuals.This makes tagging of work experience unnecessary, and the search also takes into account the background work experience of the person seeking advice and the consultation partner, as well as general-purpose knowledge of the work in question, making it possible to recommend more appropriate consultation partners.

[0160] In other words, the search system 100 acquires in the background text information of documents including business screens, case-specific documents, and manuals handled by each user terminal UT, and records and accumulates the information in association with each business practitioner (user). The search server 1 uses this accumulated information to grasp the business experience and versatile knowledge of each business practitioner, thereby eliminating the need for explicit / active registration (input, tagging, etc.) of business experience for the purpose of client pairing.

[0161] The search server 1 also divides text information related to the business information of the current consultant into "problem information TW" and "peripheral information PW." Candidates for consultation are selected based on their work experience related to the problem information TW, and a higher recommendation level is assigned to those who have more of both the problem information TW and the peripheral information PW. In this way, by using the amount of work experience possessed by each business practitioner regarding the problem information TW as an extraction condition, those other than those who have work experience related to the problem information TW are excluded from candidates for consultation. As a result, the candidates are more likely to have experience in solving problems. In addition, by determining the recommendation level of the consultation candidates using the amount of work experience possessed by each business practitioner regarding the peripheral information PW, those who have more work experience regarding the peripheral information PW are given higher priority as candidates for consultation. As a result, consultations are conducted between people with experience in the same business situation, - The person seeking advice does not need to explain a lot to the person seeking advice in order to help them understand the work situation at the time the problem occurred. - By sharing their own experiences, the person seeking advice can come up with a solution that is tailored to the person's work situation, and they do not need to prepare a separate explanation tailored to the person's work situation. etc. become possible.

[0162] Furthermore, the search server 1 calculates a set of judgment variables for the documents referenced by each of the recommendation level-attached consultation candidates, and judges whether the document is a manual or not based on this set of judgment variables. If it is a manual, it counts the number of pieces of problem information TW and peripheral information PW contained in the document as the number of manuals referenced, and assigns a priority to the recommendation level-attached consultation candidates according to the number of manuals referenced. This makes it possible to select, as a priority level consultation candidate, a person who has versatile knowledge (the meaning and reason of each task) and is likely to know how to deal with the problem even if the person himself does not handle it, rather than simply a person who has performed a series of tasks to solve the problem of the person many times, and thus increases the possibility of solving the problem.

[0163] In other words, the search server 1 can determine whether the document open on each user terminal UT was a manual or a case-specific document, thereby determining whether the business operator of the user terminal UT was viewing "manuals." In this way, the search server 1 can search for people who know the meaning and reason of the problem information TW, and can use this information to rank consultation candidates in the consultant pairing technology.

[0164] Furthermore, the search server 1 checks the possession status of the problem information TW and the work experience of all business practitioners, not just the person seeking advice. This realizes pairing that includes the perspective of the person seeking advice, not just the person seeking advice.

[0165] Furthermore, as a result of the pairing process by the search server 1, the final consultation candidate who is assigned a high priority is notified that a problem has arisen that they may be able to help with based on their own experience. This allows the business practitioner who may become the consultation candidate to easily notice the occurrence of a problem that can be solved based on their own experience and knowledge. In addition, they can easily distinguish this problem from other problems that are difficult to solve based on their own experience and knowledge.

[0166] Furthermore, from the information generated by the search server 1, in addition to the text information handled by the consultation source user terminal UTX, screen images and document paths are acquired, and are recorded and stored in association with the text information. When the final consultation candidate accepts a consultation or receives a notification, text information including the problem information TW and peripheral information PW, which are the pairing conditions, is identified from the accumulated text information, and the associated screen images and paths are displayed. If the identified text information, screen images, and paths can be shared with the person seeking advice, without the final consultation candidate having to search for past documents himself or display past cases again on the system screen, it will be possible to easily present the person seeking advice with case studies of solutions that the final consultation candidate has experienced.

[0167] Furthermore, the person seeking advice does not need to specifically input what they are having trouble with; even if they are in a situation where they "cannot specifically identify what is bothering them" or "don't know what they don't know," the system can suggest a more appropriate place to seek advice.

[0168] [Second embodiment] The same configurations and operations as those in the first embodiment are denoted by the same reference numerals as those in the first embodiment, and the description thereof will be omitted. Below, the points that differ from the first embodiment will be described.

[0169] (1) Processing Overview FIG. 18 is a schematic diagram showing an overview of the process of the search server 1 in the second embodiment.

[0170] In the first embodiment, as shown in FIG. 2, the search server 1 performs two processes, namely, "selecting business practitioners who have often used the problem information TW as consultation candidates" and "selecting, among the consultation candidates, business practitioners who have often used the peripheral information PW as consultation candidates with a high recommendation level to which they are assigned", and then performs the process of "selecting, among the consultation candidates with a recommendation level, business practitioners who have referred to manuals containing many words that constitute the business situation as prioritized consultation candidates to which a high priority level is assigned."

[0171] In contrast to this, in the second embodiment, the search server 1 performs a process of "assigning a high priority if the document containing the problem information TW is not a case-specific document but a manual" in the process of "selecting business practitioners who have often used the problem information TW as consultation candidates", and performs a process of "assigning a high priority if the document containing the peripheral information PW is not a case-specific document but a manual" in the process of "selecting business practitioners who have often used the peripheral information PW as consultation candidates with a high recommendation level to which a high recommendation level is assigned". In the second embodiment, the selected consultation candidates with a recommendation level become prioritized consultation candidates.

[0172] (2) Operation (2-1) Pairing process FIG. 19 is a diagram showing details of the process in step S4 shown in FIG.

[0173] The pairing process execution unit 31 reads out the problem information TW of the client and the peripheral information PW from the pairing condition storage unit 25 in step S41, as in the first embodiment.

[0174] Thereafter, in the second embodiment, in step S46, the pairing process execution unit 31 executes a process of extracting consultation candidates.

[0175] In step S47, the pairing process execution unit 31 executes a process of narrowing down consultation candidates.

[0176] (2-1-1) Selection of consultation candidates Fig. 20 is a flowchart showing an example of the procedure of the consultation candidate extraction process in step S46 in Fig. 19. This consultation candidate extraction process is a process in which, when "business practitioners who often use problem information TW are selected as consultation candidates", "if the document that includes the problem information TW is not a case-specific document but a manual, a high priority is assigned to it".

[0177] (2-1-1-1) Select the person to be processed In step S42, the pairing process executing unit 31 extracts service implementers who have a lot of experience using the problem information TW based on the evaluation index A, as described in the first embodiment. Then, in step S461, the pairing process execution unit 31 selects one person as a processing target from among all the extracted business practitioners.

[0178] (2-1-1-2) Documents to be processed In step S462, the pairing process execution unit 31 selects, from the terminal handling content storage unit 15, one document that the selected process target person has referred to within a predetermined period of time.

[0179] (2-1-1-3) Counting the word groups that make up the business situation In step S463, the pairing process execution unit 31 counts how many of the selected documents to be processed contain the “group of words contained in the problem information TW” that is included in the screen that the user was looking at when the problem occurred.

[0180] (2-1-1-4) Manuals Judgment In step S464, the pairing process execution unit 31 judges whether the selected document to be processed is a manual or not. As described in the first embodiment, the pairing process execution unit 31 calculates explanatory variables (group), calculates a judgment variable group from the calculated explanatory variables (group), and judges whether the document is a manual or not based on whether the calculated judgment variable group is within the judgment variable group space α of manuals that has been clustered in advance. The pairing process execution unit 31 may store the manual judgment result of each document and use it when judging the manual for other processing targets. In other words, the manual judgment by calculating the judgment variable group may be performed only for documents that have not yet been judged. If the pairing process execution unit 31 judges that the document is a manual, the process proceeds to step S465, and if the pairing process execution unit 31 judges that the document is not a manual, the process proceeds to step S466.

[0181] (2-1-1-5) Count number weighting In step S465, if the document to be processed is a manual, the pairing process execution unit 31 multiplies the number of "word groups included in the problem information TW" in the document to be processed counted in step S463 by three. Note that three times is just an example, and other values ​​may be used. If the document to be processed is not a manual, this step S465 is skipped, and the count number of "word groups included in the problem information TW" in the document to be processed remains the same as the counted number.

[0182] (2-1-1-6) Identifying potential consultants In step S466, the pairing process execution unit 31 judges whether the count number is equal to or greater than the consultation candidate discrimination threshold. The consultation candidate discrimination threshold is designated in advance by, for example, the administrator of the search server 1. If the count number is equal to or greater than the consultation candidate discrimination threshold, the pairing process execution unit 31 determines that the selected processing target is a consultation candidate, and if the count number is less than the consultation candidate discrimination threshold, the pairing process execution unit 31 does not determine that the selected processing target is a consultation candidate. If the pairing process execution unit 31 judges that the count number is equal to or greater than the consultation candidate discrimination threshold, the unit proceeds to step S467, and if the count number is less than the consultation candidate discrimination threshold, the unit proceeds to step S468.

[0183] (2-1-1-7) Save pairing results In step S467, the pairing process execution unit 31 stores the pairing result in which the processing target is a consultation candidate in the pairing result storage unit 45. For example, in the example shown in FIG. 16, the pairing process execution unit 31 stores the "consultation candidate" and the "number of references to manuals" in association with the "pairing process ID" in the pairing result storage unit 45, and can also store the "problem information TW" as an option. The "number of references to manuals" is the above-mentioned count number determined to be equal to or greater than the consultation candidate discrimination threshold, but instead of storing this count number as it is, the pairing process execution unit 31 stores a value obtained by multiplying the count number by η as a weighting factor as the "number of references to manuals". The weighting factor η is designated in advance by, for example, the administrator of the search server 1, depending on how important the "problem information TW" or the "peripheral information PW" is to be in selecting a consultation candidate.

[0184] (2-1-1-8) Update document to be processed In step S468, the pairing process execution unit 31 judges whether or not all documents referenced within a predetermined period by the business practitioner selected as the processing target in step S461 in the terminal handling content storage unit 15 have been selected as processing targets. If there are documents that have not yet been selected as processing targets, the pairing process execution unit 31 proceeds to step S462, selects the unselected documents, and repeats the above steps. On the other hand, if all documents have been selected as processing targets, the pairing process execution unit 31 proceeds to step S469.

[0185] (2-1-1-9) Update of processing target In step S469, the pairing process execution unit 31 judges whether or not all of the business practitioners other than the client have been selected as the processing targets. If there is a business practitioner who has not yet been selected as the processing target, the pairing process execution unit 31 proceeds to step S461, selects the unselected business practitioner, and repeats the above steps. If all business practitioners have been selected as the processing targets, the pairing process execution unit 31 ends the procedure of the consultation candidate extraction process in step S46. In other words, the extraction and prioritization of consultation candidates based on the problem information TW is completed.

[0186] (2-1-2) Narrow down the candidates for consultation Fig. 21 is a flow chart showing an example of the procedure of the consultation candidate narrowing down process in step S47 in Fig. 19. This consultation candidate narrowing down process is a process in which, when "among the consultation candidates to which a high priority has been assigned, a business practitioner who has used peripheral information PW a lot is selected as a consultation candidate with a high recommendation level to which a high recommendation level has been assigned," "if the document that included the peripheral information PW is not a case-specific document but a manual, a high priority is assigned."

[0187] (2-1-2-1) Recommendation Level Allocation In step S471, the pairing process execution unit 31 assigns a high recommendation level to a business practitioner who has a lot of experience with the peripheral information PW from among the consultation candidate extracted in the consultation candidate extraction process in step S46, as described in the first embodiment.

[0188] (2-1-2-2) Select the person to be processed In step S472, the pairing process executing section 31 selects one of the consultation candidates to which a high recommendation level has been assigned.

[0189] (2-1-2-3) Documents to be processed In step S473, the pairing process execution unit 31 selects, from the terminal handling content storage unit 15, one document that the selected consultation candidate to be processed has referred to within a predetermined period of time.

[0190] (2-1-2-4) Counting the word groups that make up the business situation In step S464, the pairing process execution unit 31 counts the number of times the selected document to be processed contains a "group of words contained in the peripheral information PW" that is included in the screen that the client was referring to when the problem occurred.

[0191] (2-1-2-5) Manuals Judgment In step S475, the pairing process execution unit 31 judges whether the selected document to be processed is a manual. As described in the first embodiment, the pairing process execution unit 31 calculates explanatory variable(s), calculates a judgment variable group from the calculated explanatory variable(s), and judges whether the calculated judgment variable group is within the judgment variable group space α of manuals that has been clustered in advance, thereby judging whether the document is a manual. Alternatively, the pairing process execution unit 31 may hold the manual judgment result in the above step S464 in the consultation candidate extraction process in step S46 and refer to it. If it is judged to be a manual, the process proceeds to step S476, and if it is judged not to be a manual, the process proceeds to step S477.

[0192] (2-1-2-6) Count number weighting In step S476, if the document to be processed is a manual, the pairing process execution unit 31 multiplies the number of "word groups included in the peripheral information PW" in the document to be processed counted in step S474 by three. Note that three times is just an example, and other values ​​may be used. If the document to be processed is not a manual, this step S476 is skipped, and the count number of "word groups included in the peripheral information PW" in the document to be processed remains the same as the counted number.

[0193] (2-1-2-7) Updating pairing results In step S477, the pairing process execution unit 31 updates the pairing result of the selected consultation candidate stored in the pairing result storage unit 45. For example, in the example shown in FIG. 16, the pairing process execution unit 31 updates the "number of times manuals are referred to" of the "consultation candidate" and also stores the "degree of recommendation", and can also optionally store the "peripheral information PW". When updating the "number of times manuals are referred to", the pairing process execution unit 31 does not simply add the count number, but adds a value obtained by multiplying the count number by θ as a weighting. This weighting value θ is designated in advance by, for example, the administrator of the search server 1, depending on how important the "problem information TW" or the "peripheral information PW" is to be in selecting a consultation candidate.

[0194] (2-1-2-8) Update the document to be processed In step S478, the pairing process execution unit 31 judges whether or not all documents that the consultation candidate selected in step S471 has referenced within a predetermined period of time in the terminal handling content storage unit 15 have been selected as processing targets. If there are documents that have not yet been selected as processing targets, the pairing process execution unit 31 proceeds to step S473, selects the unselected documents, and repeats the above steps. On the other hand, if all documents have been selected as processing targets, the pairing process execution unit 31 proceeds to step S479.

[0195] (2-1-2-9) Update of processing target In step S479, the pairing process execution unit 31 judges whether or not all of the consultation candidates to which a high recommendation level is assigned have been selected. If there is a consultation candidate that has not yet been selected as a processing target, the pairing process execution unit 31 proceeds to step S471, selects the unselected consultation candidate, and repeats the above steps. If all of the consultation candidates have been selected as processing targets, the pairing process execution unit 31 ends the procedure of the consultation candidate narrowing down process in step S47. That is, the narrowing down and prioritization of the consultation candidates with recommendation levels based on the peripheral information PW is completed.

[0196] As described above in detail, in the second embodiment of the present invention, in response to a search request issued by a first user (consultation source user) USX to search for candidate users from among a plurality of users, first text information handled by the terminal UTX used by the first user is acquired. Then, for words included in the first text information, the occurrence frequency in second text information acquired from a terminal device used in the past by the first user is calculated. The words included in the first text information are classified into a first word group (problem information TW) whose occurrence frequency is lower than a first threshold value and a second word group (peripheral information PW) whose occurrence frequency is equal to or higher than the first threshold value according to the occurrence frequency in the second text information. Next, for words belonging to the first word group, the occurrence frequency in third text information related to each of the past tasks of a plurality of users to be searched is determined, and candidate users are extracted according to the occurrence frequency. Furthermore, the manuals that each of the extracted candidate users has previously referred to are determined, and the priority of each candidate user is calculated so that the higher the frequency of occurrence of words belonging to the first word group in the manuals, the higher the priority. Furthermore, for the candidate users thus extracted and whose priorities have been calculated, a recommendation degree is calculated based on the frequency of occurrence of words in the second word group (peripheral information PW), and the priority of each candidate user is calculated so that the higher the frequency of occurrence of words belonging to the second word group in the manuals that each of the candidate users narrowed down based on this recommendation degree, the higher the priority. Then, information about the candidate users including the priority is presented to the first user who is the consultation source user. The consultation source user may be presented with the calculated priority or in an order according to the calculated priority. Furthermore, the candidate users may also be notified of information about the consultation source user, the problem information TW, or the peripheral information PW.

[0197] (3) Effects As described above in detail, in the second embodiment, as in the first embodiment, it is possible to search for and present to the person in charge consultation contacts who have experience with the problem information TW and the peripheral information PW and who have frequently referred to related manuals.

[0198] Furthermore, in this embodiment, when the document containing the problem information TW is not a case-specific document but a manual, the count number of "words contained in the problem information TW" in the document saved as the number of manuals referenced is, for example, tripled and saved. As a result, among the consultation candidates, those with "large sums of the word group usage frequency histories" are ranked higher as those who have viewed many manuals. In other words, since at least those who do not know the problem information TW cannot contribute to solving the problem, the consultation candidates who can become the final consultant can be those who are ranked higher at this point.

[0199] Furthermore, in this embodiment, weighting (η, θ) is performed taking into consideration the importance of the "problem information TW" and the "peripheral information PW". Therefore, by adjusting the weighting value, it is possible to change the importance of each. For example, if a person seeking advice cannot find a suitable candidate, the weighting value can be adjusted by the administrator of the search server 1 or the user terminal UTX of the person seeking advice, thereby improving the possibility of finding a candidate who can solve the problem.

[0200] [Other embodiments] It should be noted that the present invention is not limited to the above-described embodiment. For example, the search server 1 has been described as including the terminal handling content storage unit 15, the pairing condition storage unit 25, the manual-type judgment variable group space storage unit 35, and the pairing result storage unit 45, but is not limited to this. At least one of the terminal handling content storage unit 15, the pairing condition storage unit 25, the manual-type judgment variable group space storage unit 35, and the pairing result storage unit 45 may be provided in an external device separate from the search server 1. The external device may be a database server separate from the search server 1, a database provided by the cloud, or any one of the user terminals UT. Also, they may be provided in a distributed manner among these.

[0201] Similarly, the functional units F1 to F4 of the search server 1, or the terminal handling details acquisition unit 11, the pairing condition generation unit 21, the pairing condition editing unit 22, the pairing process execution unit 31, the pairing result presentation unit 41, and the pairing result notification unit 42 may be distributed and arranged in a plurality of devices, and these devices may perform processing by coordinating with each other. Also, each of these units may be realized by using a circuit. The circuit may be a dedicated circuit that realizes a specific function, or may be a general-purpose circuit such as a processor.

[0202] Also, the search server 1 has been described as extracting candidates by determining the work experience of the business practitioner to be searched for in both the problem information TW and the peripheral information PW. However, the work experience may be determined for only one of the problem information TW or the peripheral information PW. This can reduce the processing load on the search server 1 and can also increase the search processing speed. The user may be allowed to select at any time whether to use the problem information TW or the peripheral information PW for the search. In this case, a search can be performed according to the situation and preferences of the user.

[0203] Also, the search server 1 has been described as determining whether or not the user has work experience related to the problem and calculating the recommendation level based on the degree of coverage and the degree of handling as the amount of work experience possessed. However, the amount of work experience possessed may simply be the total number of occurrences of vocabulary, rather than the degree of coverage or the degree of handling. In this case, the search server 1 can select consultation candidates based on the total number of occurrences of vocabulary contained in the problem information TW, narrow down the candidates with recommendation levels based on the total number of occurrences of vocabulary contained in the peripheral information PW, and assign priorities according to the degree of experience of referring to manuals. This reduces the amount of calculation and speeds up processing.

[0204] Also, calculation of the recommendation degree may be omitted. That is, the search server 1 may select a consultation candidate based on the amount of work experience possessed and the degree of experience of referring to manuals.

[0205] Also, whether or not a document is a manual may be determined in advance in the terminal handling contents acquisition unit 11 instead of the pairing process execution unit 31. The terminal handling contents acquisition unit 11 determines whether or not a document used by the user terminal UT is a manual using a method similar to the determination method used by the pairing process execution unit 31, and stores the determination result in the terminal handling contents storage unit 15. For example, if the data held in the terminal handling contents storage unit 15 is as shown in FIG. 9, an item for "manuals" is added, and the determination result as to whether or not the document is a manual is stored there. Alternatively, a list of documents determined to be manuals may be stored in a storage unit other than the terminal handling contents storage unit 15 together with the path of the document.

[0206] In addition, in the generation process of pairing conditions described with reference to Fig. 10, the frequency of appearance of each word in the first text information corresponding to the current work situation of the consultant may be taken into consideration. For example, when a specific word appears frequently in the first text information, a process may be performed in which the amount of work experience related to the word (degree of coverage and degree of handling) strongly contributes to the degree of recommendation.

[0207] Furthermore, the flow of each process described above is not limited to the procedure described above, and the order of some steps may be changed, or some steps may be performed simultaneously in parallel. Also, the series of processes described above do not need to be performed consecutively in time, and each step may be performed at any timing.

[0208] The above-described method can be stored as a program (software means) that can be executed by a calculator (computer) on a recording medium (storage medium) such as a magnetic disk (floppy disk, hard disk, etc.), an optical disk (CD-ROM, DVD, MO, etc.), or a semiconductor memory (ROM, RAM, flash memory, etc.), and can also be distributed by transmitting it via a communication medium. The program stored on the medium side also includes a setting program that configures the software means (including not only execution programs but also tables and data structures) that the calculator executes within the computer. The computer that realizes the above-described device reads the program recorded on the recording medium, and in some cases, configures the software means by the setting program, and executes the above-described processing by controlling the operation of the software means. The recording medium in this specification is not limited to a recording medium for distribution, but also includes a storage medium such as a magnetic disk or semiconductor memory provided inside the computer or in a device connected via a network.

[0209] In addition, the processes for obtaining and dividing text information can be modified in various ways without departing from the spirit and scope of the present invention.

[0210] It should be noted that this invention is not limited to the above-mentioned embodiment, and various modifications can be made in the implementation stage without departing from the gist of the invention. The embodiments may be implemented in appropriate combination, and in that case, the combined effect can be obtained. Furthermore, the above-mentioned embodiment includes various inventions, and various inventions can be extracted by combinations selected from the multiple constituent elements disclosed. For example, if the problem can be solved and the effect can be obtained even if some constituent elements are deleted from all the constituent elements shown in the embodiment, the configuration from which the constituent elements are deleted can be extracted as an invention. [Explanation of symbols]

[0211] 1...Search server 11...Terminal handling details acquisition section 15...Terminal handling content storage section 21... Pairing condition generation unit 22…Pairing Conditions Editorial Department 25…Pairing condition storage unit 31... Pairing process execution unit 35...Manual-type judgment variable group space storage unit 41…Pairing result display section 42…Pairing result notification section 45…Pairing result storage unit 100…Search system

Claims

1. an information acquisition unit that acquires first text information handled in a terminal device used by a first user in response to a search request issued by the first user to search for candidate users from among a plurality of users; a condition generating unit that classifies words included in the first text information into a first word group whose occurrence frequency is lower than a first threshold value and a second word group whose occurrence frequency is equal to or higher than the first threshold value according to an occurrence frequency in second text information related to a past work of the first user; extracting the candidate users from the plurality of users according to an appearance frequency of a word belonging to the first word group in third text information related to a past work of each user; determining manuals that each of the extracted candidate users has previously referred to, and calculating a priority of each candidate user such that the higher the frequency of appearance of a word belonging to the first word group in the manuals, the higher the priority of the candidate user; A search processing unit; a result output unit that outputs information for presenting information about the candidate users, including the priority, to the first user; An information processing device comprising:

2. The search processing unit: Further narrowing down the candidate users from the extracted candidate users according to an appearance frequency of a word belonging to the second word group in the third text information; calculating the priority for each of the narrowed-down candidate users such that the priority is higher as the frequency of occurrence of words belonging to the first and second word groups in manuals previously referred to by each of the narrowed-down candidate users increases; the result output unit outputs, as the information regarding the candidate user, information for identifying the candidate user and the priority calculated for the candidate user. The information processing device according to claim 1 .

3. The search processing unit further narrowing down the candidate users from the extracted candidate users according to an appearance frequency of a word belonging to the second word group in the third text information; calculating a priority for each of the narrowed-down candidate users such that the higher the frequency of occurrence of a word belonging to the second word group in manuals that each of the narrowed-down candidate users has in the past, the higher the priority; Calculating a final priority based on the priority based on the occurrence frequency of words belonging to the first word group and the priority based on the occurrence frequency of words belonging to the second word group; the result output unit outputs, as the information regarding the candidate user, information for identifying the candidate user and the final priority calculated for the candidate user. The information processing device according to claim 1 .

4. The information processing device according to claim 3 , wherein the search processing unit calculates the final priority by weighting each of the priority based on the frequency of occurrence of words belonging to the first word group and the priority based on the frequency of occurrence of words belonging to the second word group.

5. The search processing unit: Conformance rate of the document with the manual format, The scope of users of the document file; The number, duration and periodicity of use of the document file a judgment variable of the document calculated based on the above-mentioned, with a judgment variable group space of manuals calculated based on a distribution of judgment variables calculated in advance from a plurality of manuals, thereby determining manuals that each of the candidate users has previously referred to; The information processing device according to claim 1 .

6. The information processing device according to claim 1 , wherein the result output unit further outputs information for presenting the information about the first user and the information about the first word group to each of the candidate users.

7. acquiring first text information handled by a terminal device used by a first user in response to a search request issued by the first user for searching for a candidate user from among a plurality of users; Classifying words included in the first text information into a first word group whose occurrence frequency is lower than a first threshold value and a second word group whose occurrence frequency is equal to or higher than the first threshold value according to an occurrence frequency in second text information related to a past work of the first user; extracting the candidate users from the plurality of users according to an appearance frequency of a word belonging to the first word group in third text information related to a past work of each user; determining manuals that each of the extracted candidate users has previously referred to, and calculating a priority of each candidate user such that the higher the frequency of appearance of a word belonging to the first word group in the manuals, the higher the priority of the candidate user; outputting information for presenting information about the candidate users, including the priority, to the first user; An information processing method executed by an information processing device.

8. A program for causing a computer to execute processing by each unit of the device according to any one of claims 1 to 6.

Images