Information processing device and information processing method

Abstract

One aspect of the present invention involves first using a frequency at which a word related to a first trouble of a first user seeking consultation appears in each text data related to past work of a plurality of other second users to select a plurality of first candidates as candidates for consultation from among the plurality of second users, and calculating, for each of the selected first candidates, a degree of recommendation on the basis of a degree of coverage or treatment of the word related to the first trouble in the text data. One aspect of the present invention involves next calculating, for each of the first candidates, a degree of similarity between a word related to a second trouble that the first candidate consulted about in the past and the word related to the first trouble of the first user seeking consultation, and correcting the degree of recommendation for each of the first candidates who resolved the second trouble that the first candidates consulted about in the past such that the degree of recommendation becomes higher as the degree of similarity becomes higher. One aspect of the present invention involves then selecting a second candidate as the candidate for consultation from among the first candidates on the basis of the corrected degree of recommendation.

Description

Information processing device and information processing method

[0001] One aspect of the present invention relates to an information processing device and an information processing method that are used to select a candidate for consultation when a user consults another user about a work-related problem while working remotely.

[0002] In remote work, it is difficult to have natural conversations between employees and other users compared to face-to-face work, and there are also fewer opportunities to see and hear about other users' work. As a result, when a work-related problem arises, it is difficult for users to determine who has the knowledge to solve the problem, and they have to spend time searching for someone to consult, which is one of the causes of a decline in work efficiency and quality.

[0003] Therefore, for example, a technology has been proposed in which the frequency of appearance of a character string in data displayed by a user on a terminal screen is regarded as an index for evaluating the user's work experience, and the recommendation level of the user as a person to consult is calculated, and users with a high recommendation level are selected as candidates to consult and paired with the person to consult (see, for example, Patent Document 1).

[0004] In the technology described in Patent Document 1, the recommendation level is calculated by calculating the recommendation level for screen data handled by multiple users, for example, by calculating the degree of coverage and degree of handling of trouble words entered by the person seeking advice, and then aggregating this recommendation level for each user to select a consultation destination.

[0005] International Publication No. 2022 / 091281

[0006] However, in the method of calculating the recommendation level described in Patent Document 1, the same problem words tend to be selected for multiple problems with similar content. As a result, a high recommendation level is calculated for the same user, which makes it more likely that a specific user will be selected as a candidate for consultation for multiple problems with similar content. As a result, the work of dealing with the problems is concentrated on a specific user, which may increase the user's workload and reduce work efficiency. Furthermore, for users who request advice, the concentration of workload on the specific user selected as a candidate for consultation means that they have to wait a long time before they can get advice, which makes it take longer to resolve their problem.

[0007] This invention has been made in light of the above circumstances, and aims to provide a technology that prevents consultations about problems from concentrating on specific users, thereby reducing the burden on users who are the subject of consultations and shortening the waiting time required for consultation.

[0008] To solve the above problem, a first aspect of an information processing device or method according to the present invention performs a process of selecting a candidate consultation destination from among a plurality of second users in response to a request for selection of a consultation destination for a first problem of a first user requesting consultation, first selecting a plurality of first candidates as candidate consultation destinations from among the plurality of second users based on the frequency of appearance of words related to the first problem of the first user in text data related to past tasks of the plurality of second users, and calculating a recommendation degree for each of the selected first candidates based on the coverage or usage of words related to the first problem in the text data. Next, for each of the first candidates, a similarity between words related to a second problem that the first candidate previously consulted about as a consultant and words related to the first problem of the first user is calculated, and the recommendation degree is corrected for first candidates that have solved the second problem previously consulted about to a value that increases as the similarity increases. Then, based on the corrected recommendation level, a second candidate that is more appropriate as a consultation destination is selected from the first candidates.

[0009] According to a first aspect of the present invention, among a plurality of first candidates selected as a person to consult based on the frequency of appearance of words related to the problem, the recommendation degree of a candidate who has previously been the source of advice and had the problem solved is corrected based on the similarity of the words related to the problem so that the recommendation degree becomes higher the higher the similarity, and a candidate to consult is selected based on this corrected recommendation degree.

[0010] In other words, among the multiple first candidates selected as a person to consult with, a candidate who is likely to have consulted about a similar problem in the past and who has had the experience of solving the problem through that consultation is more likely to be selected as a person to consult with.

[0011] Therefore, the concentration of consultations on a specific user about a problem is reduced, which reduces the burden on the consultation response work of the specific user and makes it possible to prevent a decline in work efficiency. Also, for the user who is making a consultation, the waiting time for consultation is shortened by dispersing the potential users to consult with, which makes it possible to solve the problem quickly.

[0012] In a second aspect of the present invention, the number of times each of the first candidates has been selected as the first candidate in the past for a second problem whose similarity is equal to or greater than a predetermined threshold is calculated for each of the first candidates, and the recommendation degree for each of the first candidates is corrected to a value that decreases as the number of times increases, and the corrected recommendation degree is used to select a consultation candidate.

[0013] According to a second aspect of the present invention, the recommendation level of a first candidate is corrected in accordance with the number of times the first candidate has been selected as a consultation candidate in the past for a second problem similar to the first problem of the user who is the consultation source, so that the recommendation level decreases as the number of times increases. In other words, the recommendation level of a consultation candidate who has received many consultations about similar problems in the past is corrected to decrease.

[0014] Therefore, the probability that a specific user will be selected as a consultation candidate is further reduced, making it possible to more effectively prevent consultation response work from being concentrated on a specific user. The above effects contribute to improving work efficiency, especially in remote environments where opportunities for communication between users are likely to be limited.

[0015] Furthermore, according to the first and second aspects, it is possible to increase and diversify the number of users who can be candidates for consultation about problems, thereby contributing to the efficient dissemination of business knowledge and skills throughout the organization.

[0016] In other words, according to one aspect of the present invention, it is possible to provide a technology that prevents consultations regarding problems from concentrating on specific users, thereby reducing the burden on the users being consulted and shortening the waiting time required before receiving a consultation.

[0017] FIG. 1 is a diagram illustrating an example of the configuration of a system that realizes a consultation candidate selection service according to an embodiment of the present invention. FIG. 2 is a block diagram illustrating an example of the hardware configuration of a server device that constitutes the core of the system illustrated in FIG. 1. FIG. 3 is a block diagram illustrating an example of the software configuration of the server device that constitutes the core of the system illustrated in FIG. 1. FIG. 4 is a flowchart illustrating an example of the overall processing procedure and processing content of a consultation candidate selection process executed by a control unit of the server device illustrated in FIG. 3. FIG. 5 is a flowchart illustrating an example of the processing procedure and processing content of a pairing process among the processing procedures illustrated in FIG. 4. FIG. 6 is a flowchart illustrating an example of the processing procedure and processing content of a process for calculating similarity between problem words among the processing procedures illustrated in FIG. 4. FIG. 7 is a flowchart illustrating an example of the processing procedure and processing content of a first recommendation degree correction process among the processing procedures illustrated in FIG. 4. FIG. 8 is a flowchart illustrating an example of the processing procedure and processing content of a second recommendation degree correction process among the processing procedures illustrated in FIG. 4. FIG. 9 is a diagram illustrating an example of a user who requests consultation and multiple users who are candidate consultation users. FIG. 10 is a diagram illustrating an example of problem words included in text data handled by the user who requests consultation and multiple users who are candidate consultation users. Fig. 11 is a diagram showing an example of a consultation candidate list obtained by the pairing process in the processing procedure shown in Fig. 4, an example of problem words that consultation users included in the consultation candidate list have dealt with in the past, and an example of a calculation result of the similarity between the problem words. Fig. 12 is a diagram showing an example of a representative similarity selected from the similarities between the problem words shown in Fig. 11. Fig. 13 is a diagram showing an example of a correction result of the first recommendation degree. Fig. 14 is a diagram showing an example of a correction result of the second recommendation degree.

[0018] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0019] [One Embodiment] (Configuration Example) (1) System FIG. 1 is a diagram showing an example of the configuration of a system that realizes a consultation candidate selection service according to one embodiment of the present invention.

[0020] The system of one embodiment includes a server device SV that executes a consultation candidate selection process, and enables data communication via a network NW between this server device SV and multiple user terminals UT1 to UTn used by users such as employees.

[0021] The user terminals UT1 to UTn are, for example, personal computers equipped with a browser or a mailer. Note that the user terminals UT1 to UTn may also be mobile terminals such as smartphones and tablet terminals.

[0022] The network NW comprises, for example, a wide area network centered on the Internet and an access network for accessing this wide area network. Examples of the access network include, but are not limited to, a public communication network using wired or wireless connections, a local area network (LAN) using wired or wireless connections, and a cable television (CATV) network. Note that, if the service provision area is limited to a company's business establishment or office, the network NW may be configured solely with a LAN or wireless LAN.

[0023] (2) Server SV The server SV is configured by a server computer installed on the cloud or the web, for example. Note that the server SV may be configured by a personal computer used by a system administrator, for example.

[0024] 2 and 3 are block diagrams showing an example of the hardware configuration and software configuration of the server device SV, respectively.

[0025] The server device SV has a control unit 1 that uses a hardware processor such as a central processing unit (CPU), and this control unit 1 is connected via a bus 5 to a storage unit having a program storage unit 2 and a data storage unit 3, and a communication interface (hereinafter, the interface will be referred to as I / F) unit 4.

[0026] The communication I / F unit 4 transmits and receives data to and from the user terminals UT1 to UTn using a communication protocol defined in the network NW. The communication I / F unit 4 also communicates data with an administrator terminal (not shown) used by a system administrator or the like.

[0027] The program storage unit 2 is configured by combining, for example, a nonvolatile memory such as a solid state drive (SSD) as a storage medium that can be written to and read from at any time, and a nonvolatile memory such as a read only memory (ROM), and stores middleware such as an operating system (OS) as well as application programs required to execute various controls according to the first embodiment. Hereinafter, the OS and each application program will be collectively referred to as a program.

[0028] The data storage unit 3 is, for example, a combination of a non-volatile memory such as an SSD that can be written to and read from at any time, and a volatile memory such as a RAM (Random Access Memory), and the storage area includes a terminal handling data storage unit 31, a pairing condition storage unit 32, a consultation candidate list storage unit 33, and a consultation history storage unit 34.

[0029] The terminal handling data storage unit 31 is used to store handling data acquired from the user terminals UT1 to UTn for each user. The handling data includes text data handled by the user terminals UT1 to UTn, as well as displayed screen images and paths to documents being used. The handling data may also include voice data input by a microphone, for example.

[0030] The pairing condition storage unit 32 is used to store a group of problem words and a group of related words that are classified and extracted from the handling data of the user (also called the client) who is the source of the consultation, in association with the client's identification information.

[0031] The consultation candidate list storage unit 33 stores a list of consultation candidates selected by a pairing process (described later) in association with the identification information of the person seeking advice. The consultation candidate list includes information on a plurality of consultation candidates. Each piece of consultation candidate information is composed of a pair of identification information of a user who is a consultation candidate and a recommendation level of the user.

[0032] The consultation history storage unit 34 stores information indicating the content of consultations that have been held in the past as consultation history information. The consultation history information includes, for example, the person who made the consultation, potential consultation partners, problem words, and information indicating whether the consultation was successful or not. The information indicating whether the consultation was successful or not is information indicating whether the problem was solved by the consultation, and is registered, for example, by the person who made the consultation.

[0033] The control unit 1 includes, as processing functions necessary to implement one embodiment of the present invention, a terminal handling data acquisition processing unit 11, a pairing condition generation processing unit 12, a pairing execution processing unit 13, a similarity calculation processing unit 14, a first recommendation degree correction processing unit 15, a second recommendation degree correction processing unit 16, a consultation candidate selection processing unit 17, and a selection information output processing unit 18.

[0034] The processing units 11 to 18 are all realized by causing a hardware processor in the control unit 1 to execute an application program stored in the program storage unit 2. Note that some or all of the processing units 11 to 18 may be realized using hardware such as an LSI (Large Scale Integration) or an ASIC (Application Specific Integrated Circuit).

[0035] The terminal handling data acquisition processing unit 11 acquires data handled by the user terminal UTi of the user (consultant) who has encountered a problem (first problem) from the user terminal UTi via the communication I / F unit 4. Then, the acquired text data of the handling data is stored in the terminal handling data storage unit 31 in association with the identification information of the consultant.

[0036] The pairing condition generation processing unit 12 divides text data acquired from the user terminal UTi of the client when a problem occurs into vocabulary words, and calculates the frequency of occurrence of each divided vocabulary word in text data acquired during past consultations by the client.The pairing condition generation processing unit 12 then determines whether each vocabulary word corresponds to a "problem word" or a "related word" based on the calculated frequency of occurrence of the vocabulary word, and stores the "problem word group" and "related word group" classified based on this determination result in the pairing condition storage unit 32 in association with the client's identification information.An example of the pairing condition generation processing will be described in the operation example.

[0037] The pairing execution processing unit 13 compares the "group of problem words" and the "group of related words" stored in the pairing condition storage unit 32 with the text data of a plurality of other users stored in the terminal handling data storage unit 31. Then, based on the comparison results, a plurality of users who have a lot of experience using the "group of problem words" and the "group of related words" are selected from the other plurality of users as consultation candidates.

[0038] In addition, for each of the selected users who are candidates for consultation, the pairing execution processing unit 13 calculates the degree of coverage or the degree of handling of each of the ``group of problem words'' and ``group of related words'' in the text data of the user as the amount of work experience the user has regarding the problem and related situations, and calculates a recommendation level (first recommendation level) based on the calculated degree of coverage or degree of handling.

[0039] Then, the pairing execution processing unit 13 stores a set of user identification information and the recommendation level for each of the plurality of users selected as consultation candidates in the consultation candidate list storage unit 33 as consultation candidate information.

[0040] The similarity calculation processing unit 14 reads, for each of the users who are candidates for consultation selected by the pairing execution processing unit 13, from the consultation history storage unit 34, a group of problem words (second problem) extracted when the user of the candidate for consultation was previously selected by the pairing processing, and calculates the similarity between the read group of problem words and the group of problem words of the client (first problem) for each problem word. Then, the similarity calculation processing unit 14 selects a representative similarity from the calculated similarities for each problem word.

[0041] The first recommendation degree correction processing unit 15 targets users who have solved a problem in a past consultation among the users who are candidates for consultation included in the list of candidate consultations, and corrects the first recommendation degree of each user based on the value of the representative similarity calculated for that user so that the higher the representative similarity, the higher the first recommendation degree. Note that an example of this first recommendation degree correction processing will be described in the operation example.

[0042] The second recommendation degree correction processing unit 16 calculates, for each user of the consultation candidate list, the number of times that the user has been selected as a consultation candidate in the past for problems for which the representative similarity is equal to or greater than a threshold value (number of consultations handled), from the consultation history information of the corresponding user stored in the consultation history storage unit 34. Then, the second recommendation degree correction processing unit 16 corrects the first recommendation degree of the user in accordance with the calculated number of consultations handled so that the first recommendation degree decreases as the number of consultations handled increases. Note that an example of this second recommendation degree correction processing will also be described in an operation example.

[0043] The consultation candidate selection processing unit 17 selects a predetermined number of users in descending order of the second recommendation degree and the third recommendation degree from the multiple users selected as consultation candidates by the pairing execution processing unit 13, based on the recommendation degree (second recommendation degree) corrected by the first recommendation degree correction processing unit 15 and the recommendation degree (third recommendation degree) corrected by the second recommendation degree correction processing unit 16. Then, the consultation candidate selection processing unit 17 notifies the selected consultation candidate users to the selection information output processing unit 18 and stores them in the consultation history storage unit 34.

[0044] The selected information output processing unit 18 generates presentation information regarding the users of the consultation candidate selected by the consultation candidate selection processing unit 17, and transmits the generated presentation information from the communication I / F unit 4 to the user terminal UTi of the person seeking advice.

[0045] (Example of Operation) Next, an example of operation of the server device SV configured as above will be described.

[0046] Here, as shown in FIG. 9, an example will be described in which an employee A has a problem and requests the server device SV to select a candidate for consultation.

[0047] (I) Management of Preparation Data The terminal handling data storage unit 31 in the data storage unit 3 stores text data of handling data acquired from the user terminals UT1 to UTn of each employee (user) who has previously consulted about a problem, linked to the employee information and employee identification information (user ID) of the employee who made the consultation. The employee information includes employee attribute information and address information of the user terminals UT1 to UTn used, etc.

[0048] Information indicating the content of each consultation that has been held in the past is stored as consultation history information in the consultation history storage unit 34. Each piece of consultation history information includes, for example, identification information of the person making the consultation, a list of employees who are candidates for consultation, words used to describe the problem that was the subject of the consultation, and consultation success / failure information indicating whether the problem was resolved through the consultation.

[0049] Although not shown, the data storage unit 3 stores a unit correction amount that the second recommendation degree correction processing unit 16 uses when correcting the recommendation degree of an employee who is a consultation candidate. This unit correction amount is the correction amount per number of times an employee has been selected as a consultation candidate in the past (hereinafter also referred to as the number of consultations), and is set individually for each employee.

[0050] (II) Selection of Consultation Candidates FIG. 4 is a flowchart showing an example of the procedure and content of a consultation candidate selection process executed by the control unit 1 of the server device SV.

[0051] (1) Acquisition of the client's terminal operating data When a request to select a candidate for consultation is sent from the user terminal UTi of employee A, the client, the control unit 1 of the server device SV first executes the terminal operating data acquisition process as follows in step S1 under the control of the terminal operating data acquisition processing unit 11.

[0052] That is, in response to the selection request, the terminal operation data acquisition processing unit 11 acquires terminal operation data including screen data and the like used on employee A's user terminal UTi, i.e., terminal operation data at the time of the problem occurrence, from the user terminal UTi via the communication I / F unit 4. Then, the text data of the acquired terminal operation data is stored in the terminal operation data storage unit 31 in association with the client's identification information (client ID). Note that if the terminal operation data includes non-text data such as voice data, this data is converted to text data.

[0053] (2) Pairing Process Next, in step S2, the control unit 1 of the server device SV, under the control of the pairing condition generation processing unit 12 and the pairing execution processing unit 13, executes a process for pairing the above-mentioned client with a consultation candidate as follows.

[0054] FIG. 5 is a flowchart showing an example of the processing procedure and processing content of the pairing condition generation processing and pairing execution processing executed by the pairing condition generation processing unit 12 and the pairing execution processing unit 13.

[0055] (2-1) Generation of Pairing Conditions First, in step S20, under the control of the pairing condition generation processing unit 12, the control unit 1 of the server device SV reads out the text data acquired when the problem occurred to the employee A who is seeking advice from the terminal handling data storage unit 31, and divides the read out text data into vocabulary using, for example, morphological analysis.

[0056] Next, in step S21, the pairing condition generation processing unit 12 calculates the frequency of appearance of each of the divided vocabulary words in the text data of past consultations of employee A stored in the terminal handling data storage unit 31. The "frequency of appearance" is calculated by dividing the number of times that each word included in the text data when the problem occurred appears in the past text data to be compared by the total number of all words included in the text data to be compared.

[0057] In step S22, the pairing condition generation processing unit 12 compares the obtained frequency of occurrence of the above vocabulary with a preset threshold value and divides the vocabulary into a first group of words whose frequency of occurrence is lower than the threshold value and a second group of words whose frequency of occurrence is equal to or higher than the threshold value. The first group of words is then defined as a "problem word group" that includes words that the client has less experience with, while the second group of words is defined as a "peripheral word group."

[0058] As a method for determining the threshold value, for example, a method is used in which statistical values ​​(average value, median value, mode value, variance, standard deviation, etc.) of the frequency of occurrence of vocabulary in the text data accumulated by the client himself / herself are calculated and the threshold value is determined from the statistical values, or a method is used in which statistical values ​​(average value, median value, mode value, variance, standard deviation, etc.) of the frequency of occurrence of vocabulary in the text data accumulated by all users are calculated and the threshold value is determined from the statistical values, but a method in which the system administrator arbitrarily sets the threshold value as a fixed value may also be used.

[0059] When the vocabulary classification process is completed, the pairing condition generation processing unit 12 stores the classified "trouble word group" and "related word group" in the pairing condition storage unit 32 in association with the client ID.

[0060] In addition, in step S23, the pairing condition generation processing unit 12 transmits the classified "group of problem words" and "group of related words" from the communication I / F unit 4 to the user terminal UTi of the person seeking advice, employee A, and displays them on the screen.

[0061] In response, employee A determines whether the displayed "group of problem words" and "group of related words" correspond to words that describe his / her current problem or words that describe the surrounding situation, and performs an editing operation to delete, change, or add words as necessary. This editing operation information is transmitted from the user terminal UTi to the server device SV.

[0062] In step S24, the pairing condition generation processing unit 12 acquires the editing operation information transmitted from the user terminal UTi via the communication I / F unit 4, and edits the "group of problem words" and "group of related words" stored in the pairing condition memory unit 32 in accordance with the acquired editing operation information.

[0063] (2-2) Execution of Pairing Next, under the control of the pairing execution processing unit 13, the control unit 1 of the server device SV executes a process of searching for other employees who are candidates for consultation with employee A, the person seeking advice, as follows.

[0064] That is, in step 25, the pairing execution processing unit 13 first reads text data of past consultations acquired from multiple employees other than the person seeking advice, employee A, from the terminal handling data storage unit 31. For example, Fig. 10 shows a case where text data including past trouble words is read from employees B, C, and X.

[0065] Then, in step S26, the pairing execution processing unit 13 calculates the frequency of appearance of the "trouble word group" and the "related word group" in the text data of the other employees that has been read in. Next, in step S27, the pairing execution processing unit 13 selects a plurality of candidates (first candidate users) to be consulted from among the other employees based on the frequency of appearance of the "trouble word group" and the "related word group."

[0066] Furthermore, in step S28, the pairing execution processing unit 13 calculates a recommendation level (first recommendation level) for each of the other employees selected as the first candidate users as a person to consult with.

[0067] In other words, the process of selecting the candidate consultation destination and the process of calculating the first recommendation level are performed by calculating the degree of coverage or handling of each of the problem words (also called problem information) and related word groups (also called related information) in the text data at the time of each employee's consultation as the amount of work experience that each employee has regarding the problem and related circumstances.

[0068] Finally, in step S29, the pairing execution processing unit 13 stores a pair of the selected first candidate user and the first recommendation degree calculated for the candidate user in the consultation candidate list storage unit 33, including the pair in the consultation candidate list information.

[0069] (2-3) Detailed Example of Pairing Process An example of the above-described process of selecting a first candidate user and the process of calculating a first recommendation degree will be described in more detail. In the following description, the following are defined: A set of all task performing users (including the client and the client): K A set of records for each task performing user k (∈K): Lk A set of problem words (a set of vocabulary words included in the set): T A set of peripheral words (a set of vocabulary words included in the set): B

[0070] The pairing execution processing unit 13 performs pairing for each word t included in the problem word group T. ｉ Regarding the vocabulary, the occurrence frequency of the corresponding word: m ｉ，ｋ Then, calculate the following as the "amount of work experience related to the group of problem words."

[0071]

[0072] The coverage of problem information, in other words, represents the proportion of words contained in the problem word group that appear in each employee's record. The coverage of problem information, in other words, represents the sum of the frequencies at which words belonging to the problem word group appear in each employee's record divided by the total number of words belonging to the problem word group.

[0073] The pairing execution processing unit 13 also performs pairing for each word b included in the surrounding word group B. ｊ From the vocabulary frequency distribution, the occurrence frequency of the corresponding word n ｊ，ｋThen, calculate the following as the "amount of work experience related to peripheral information."

[0074]

[0075] The coverage of peripheral information indicates the proportion of words in the peripheral word group that appear in each employee's record. The handling of peripheral information indicates the sum of the frequencies at which words in the peripheral word group appear in each employee's record divided by the total number of words in the peripheral word group.

[0076] Furthermore, the pairing execution processing unit 13 extracts a plurality of consultation candidates from the set K using the "amount of work experience related to the problem information," and for each of the extracted consultation candidates k, calculates the following "recommendation level" r using the "amount of work experience related to the problem information" and the "amount of work experience related to the peripheral information." ｋ is calculated and assigned to each consultation candidate k.

[0077]

[0078] In the above formula, α, β, γ, and δ are parameters that control which of the evaluation indicators related to the “amount of work experience related to problem information” and the “amount of work experience related to peripheral information” should be reflected in the recommendation level, and to what extent, and are specified in advance by, for example, an administrator of the server device SV.

[0079] In other words, the first recommendation level is a value obtained by weighting the degree of coverage of problem information, the degree of handling of problem information, the degree of coverage of peripheral information, and the degree of handling of peripheral information for each consultation candidate k. The first recommendation level takes a higher value as the value of the degree of coverage of problem information, the degree of handling of problem information, the degree of coverage of peripheral information, and the degree of handling of peripheral information become larger.

[0080] In addition, the calculation of the first recommendation level does not need to use all of the degree of coverage of problem information, the degree of handling of problem information, the degree of coverage of peripheral information, and the degree of handling of peripheral information, but may use any one of these. In other words, the above α, β, γ, and δ may be positive values ​​or zero, and it is sufficient that at least one of α, β, γ, and δ is non-zero.

[0081] (3) Calculation of Similarity Between Problem Words Next, in step S3, under the control of the similarity calculation processing unit 14, the control unit 1 of the server device SV executes a process of calculating the similarity between a group of problem words in the past consultation of the first candidate user and a group of problem words of the client as follows.

[0082] FIG. 6 is a flowchart showing an example of the procedure and content of the similarity calculation process executed by the similarity calculation processing unit 14.

[0083] That is, first, in step S30, the similarity calculation processing unit 14 selects one first candidate user from the consultation candidate list storage unit 33. Then, in step S31, the similarity calculation processing unit 14 reads a group of problem words that the selected first candidate user has consulted about in the past from the consultation history storage unit 34. Subsequently, in step S32, the similarity calculation processing unit 14 calculates the similarity between the group of problem words of the first candidate user and the group of current problem words of the person seeking advice, for each problem word.

[0084] Next, in step S33, the similarity calculation processing unit 14 selects a representative similarity from the similarities calculated for each problem in the past consultations of the candidate user.

[0085] When the similarity calculation and representative similarity selection process for one candidate user is completed, the similarity calculation processing unit 14 determines whether or not there are any unselected candidate users in step S34. If the result of this determination is that there are unselected candidate users remaining, the process returns to step S30 to select the next candidate user, and thereafter executes similarity calculation and representative similarity selection in steps S31 to S33. Thereafter, the similarity calculation and representative similarity selection process is repeatedly executed for all unselected candidate users. When the process for all candidate users is completed, the similarity calculation process ends.

[0086] Here, an example of calculating the similarity will be described. For example, suppose that employees B, C, and X shown in FIG. 10 are selected as first candidate users, and the problem in the past consultation of employee B is similar to that of employee C. 1-1 , C 1-2 , C 1-3 Employee C's past problems were 2-1 , C 2-2 Employee X's past problems were 3-1 Let's say that was the case.

[0087] The similarity calculation processing unit 14 calculates the similarity of each of the first candidate users, Employee B, Employee C, and Employee X, for each of the first candidate users, Employee C, and Employee X. 1-1 , C 1-2 , C 1-3 , C 2-1 , C 2-2 , C 3-1 A group of words representing problems W C1-1 , W C1-2 , W C1-3 , W C2-1 , W C2-2 , W C3-1 Load each of them.

[0088] For example, employee B's problem C 1-1 Regarding W C1-1 = (W C1-1-1 , W C1-1-2 , ...) as problem C 1-2 Regarding W C1-2 = (W C1-2-1 , W C1-2-2 , ...) as problem C 1-3 Regarding W C1-3 = (WC1-3-1 , W C1-3-2 , ...) are read respectively. Also, employee C's problem C 2-1 Regarding W C2-1 = (W C2-1-1 , W C2-1-2 , ...) as problem C 2-2 Regarding W C2-2 = (W C2-2-1 , W C2-2-2 Similarly, employee X's problem C 3-1 Regarding W C3-1 = (W C3-1-1 , W C3-1-2 , ...).

[0089] Then, the similarity calculation processing unit 14 calculates the similarity of each of the problems C of the employees B, C, and X that have been read. 1-1 , C 1-2 , C 1-3 , C 2-1 , C 2-2 , C 3-1 A group of words representing problems W C1-1 , W C1-2 , W C1-3 , W C2-1 , W C2-2 , W C3-1 and the client's current trouble word group W' = (W' 1 , W' 2 , ...) C1-1 , R C1-2 , R C1-3 , R C2-1 , R C2-2 、 R C3-1 Calculate.

[0090] For example, the trouble word W C1-1 The current problem of the person asking for advice is word W. 1 Similarity R C1-1 is calculated by the following formula:

[0091]

[0092] FIG. 11 shows the first candidate users, Employee B, Employee C, and Employee X, and the first candidate users ... C, and Employee X, and the first candidate users, Employee C, and Employee X, and the first candidate users, Employee C 1-1 , C 1-2 , C 1-3 , C 2-1 , C 2-2 , C3-1 And the troublesome word group W C1-1 , W C1-2 , W C1-3 , W C2-1 , W C2-2 , W C3-1 and the calculated similarity R C1-1 , R C1-2 , R C1-3 , R C2-1 , R C2-2 、 R C3-1 10 is a diagram showing an example of the first recommendation level in a table format together with an example of calculation of the first recommendation level.

[0093] The similarity calculation processing unit 14 also selects the representative similarity as follows.

[0094] That is, the similarity calculation processing unit 14 calculates the problem C for each of the employees B, C, and X. 1-1 , C 1-2 , C 1-3 , C 2-1 , C 2-2 , C 3-1 The similarity R calculated for C1-1 , R C1-2 , R C1-3 , R C2-1 , R C2-2 、 R C3-1 From among these, the representative similarity R P1 , R P2 , ... are selected. Representative similarity R P1 , R P2 , . . . , the similarity with the largest value (maximum value of similarity) is selected. P1 , R P2 , . . . are examples of selection results.

[0095] (4) First recommendation degree correction process The first recommendation degree correction process corrects the recommendation degree of employees included in the consultation candidate list generated by the pairing process who are likely to have consulted with other employees about a similar problem in the past and who have had the experience of having their problem resolved through this consultation. By performing this correction, employees who meet the above conditions are more likely to be selected as consultation candidates.

[0096] To perform the above correction, the control unit 1 of the server device SV executes the following process in step S4 under the control of the first recommendation degree correction processing unit 15.

[0097] FIG. 7 is a flowchart showing an example of the procedure and content of the correction process executed by the first recommendation degree correction processing unit 15.

[0098] That is, in step S40, the first recommendation level correction processing unit 15 first selects a candidate user who has solved the problem corresponding to the representative similarity from the list of candidate consultation destinations generated by the above-mentioned pairing process, based on information indicating the success or failure of the consultation contained in the past consultation history information corresponding to the above-mentioned candidate user stored in the consultation history storage unit 34.

[0099] 13, for example, suppose that the problem of employee B is not resolved (resolved "X"), and the problems corresponding to the representative similarities of employees C and X are resolved (resolved "O") In this case, the first recommendation degree correction processing unit 15 selects employees C and X as candidates for correction.

[0100] Next, in step S41, the first recommendation degree correction processing unit 15 calculates the average value of the representative similarities of employee C and employee X who solved the selected problem. Then, in step S42, the first recommendation degree correction processing unit 15 calculates the ratio of the representative similarity of each employee C and employee X to the calculated average representative similarity value, and in step S43, corrects the first recommendation degree of each employee C and employee X to a second recommendation degree using the calculated representative similarity ratio.

[0101] After completing the correction of the recommendation degree for one employee, the first recommendation degree correction processing unit 15 determines in step S44 whether there are any unselected employees among the employees whose problems have been resolved. If this determination determines that there are any unselected employees, the process returns to step S40, selects the next employee, and performs the correction processing in steps S41 to S43 on this employee's first recommendation degree. Thereafter, the first recommendation degree correction processing unit 15 repeatedly performs the correction processing in steps S41 to S43 on the first recommendation degrees of all unselected employees.

[0102] An example of the first recommendation degree correction process described above will be described with reference to Fig. 13. In the example shown in Fig. 13, the representative similarity R of each of employees C and X who solved the problem corresponding to the representative similarity is P2 , R P3 are "0.9" and "0.5" respectively, so the average value R avg Therefore, the first recommendation degree correction processing unit 15 calculates the average value R avg Representative similarity R of employee C to P2 The ratio (0.9 / 0.7) is multiplied by the first recommendation level of employee C (=35) to calculate a correction value of 45. Similarly, for employee X, the average value R avg Representative similarity R of employee X to P3 The ratio (0.5 / 0.7) is multiplied by the first recommendation level of employee X (70) to calculate a correction value (50).

[0103] That is, the first recommendation level of each employee is S P1 , S P2 , ..., the representative similarity is R P1 , R P2 , ..., and the average value is R avg Then, the corrected second recommendation level S' is P1 , S′ P2 , ... are calculated as follows:

[0104]

[0105] Finally, the first recommendation level correction processing unit 15 determines whether the corrected recommendation level (second recommendation level) is higher or lower than the recommendation level before correction (first recommendation level). If the result of this determination is lower, the correction is not adopted. Therefore, in the example shown in Figure 13, the correction to the first recommendation level of employee X is not adopted, and the recommendation level of employee X remains at the first recommendation level before correction = 70.

[0106] In contrast, the recommendation level after correction (second recommendation level) for employee C is higher than the recommendation level before correction (first recommendation level). Therefore, the correction is determined to be valid, and the recommendation level for employee C is corrected to "45."

[0107] That is, in the first recommendation degree correction process, the recommendation degree of each employee who solved the past problem corresponding to the representative similarity is corrected so that the recommendation degree becomes higher as the representative similarity of the problem has a higher value. Then, only the correction that results in a higher recommendation degree after correction than the recommendation degree before correction is adopted.

[0108] (5) Second recommendation degree correction process The second recommendation degree correction process calculates the number of times that another employee has previously responded to consultations regarding a problem similar to that of the employee requesting consultation.Then, in accordance with the number of times that another employee has previously responded to consultations, the first recommendation degree of the other employee is corrected so that the higher the number of times, the lower the first recommendation degree.In other words, the second recommendation degree correction process corrects the recommendation degree of an employee who has previously received many consultations regarding similar problems.

[0109] In order to perform the above correction process, the control unit 1 of the server device SV executes the following process in step S5 under the control of the second recommendation degree correction processing unit 16.

[0110] FIG. 8 is a flowchart showing an example of the procedure and content of the correction process executed by the second recommendation degree correction processor 16.

[0111] First, in step S50, the second recommendation degree correction processing unit 16 selects one employee from the list of consultation candidates generated by the pairing process described above. In step S51, for the selected employee, the second recommendation degree correction processing unit 16 selects problems for which the employee has previously been selected as a consultation candidate, with a similarity equal to or greater than a threshold value. The threshold value may be, for example, a fixed value set by the system administrator.

[0112] In addition, in step S51, the second recommendation degree correction processing unit 16 counts the number of times the employee has been selected as a candidate for consultation regarding the selected problem in the past, that is, the number of times the employee has responded to consultations regarding the problem in the past, based on the consultation history information related to the employee stored in the consultation history memory unit 34.

[0113] Next, in step S52, the second recommendation degree correction processing unit 16 corrects the first recommendation degree of the employee according to the counted number of consultations so that the more consultations the employee has handled, the lower the first recommendation degree. For example, the second recommendation degree correction processing unit 16 stores a unit correction amount per consultation in a storage area in the data storage unit 3, multiplies the number of consultations by the unit correction amount, and subtracts this correction amount from the first recommendation degree. The recommendation degree after this subtraction is then used as the corrected recommendation degree (third recommendation degree) for the employee.

[0114] After completing the correction of the recommendation level for one employee in the consultation candidate list, the second recommendation level correction processing unit 16 determines in step S53 whether there are any unselected employees in the consultation candidate list. If this determination shows that there are any unselected employees remaining, the process returns to step S50, selects the next employee, and performs the correction processing in steps S51 to S52 on the recommendation level of this employee. Thereafter, the second recommendation level correction processing unit 16 repeatedly performs the correction processing in steps S51 to S52 on the recommendation levels of all unselected employees.

[0115] An example of the second recommendation degree correction process described above will be explained with reference to Fig. 14. The second recommendation degree correction processing unit 16 corrects the problem similarity R calculated by the similarity calculation processing unit 14 for each of employees B, C, and X included in the consultation candidate list. C1-1 , R C1-2 , R C1-3 , R C2-1 , R C2-2 , R C3-1 14, the problem C with the similarity of employee B=0.7 is selected. 1-3 and employee C's problem C, whose similarity is 0.9. 2-2 and employee X's similarity to problem C, which has a similarity of 0.5. 3-1 is selected.

[0116] The second recommendation degree correction processing unit 16 searches the consultation history information about the employees B, C, and X stored in the consultation history storage unit 34 to find the problem C. 1-3, Problem C 2-2 and Problem C 3-1 The number of times that each of the above employees B, C, and X has been selected as a candidate for consultation is counted. 3-1 The number of past consultations regarding this issue is counted as "2."

[0117] Therefore, the second recommendation degree correction processing unit 16 calculates the recommendation degree of employee X (=70) as 70 - 10 x 2 = 50 based on the number of consultations handled (=2) and the unit correction amount per consultation (=10) that is uniformly set for each employee, and sets the calculated value (=50) as the new recommendation degree (third recommendation degree) of employee X.

[0118] (6) Selection of candidate consultation partners After the first and second recommendation degree correction processes described above are completed, the control unit 1 of the server device SV, under the control of the candidate consultation selection processing unit 17, selects, in step S6, employees to be presented to the employee seeking advice as candidates for consultation about their problem, as follows:

[0119] That is, based on the recommendation degree (second recommendation degree) corrected by the first recommendation degree correction processing unit 15 and the recommendation degree (third recommendation degree) corrected by the second recommendation degree correction processing unit 16, the consultation candidate selection processing unit 17 selects a predetermined number of users from the multiple users selected as consultation candidates by the pairing execution processing unit 13 in descending order of the second recommendation degree and the third recommendation degree.

[0120] For example, in the example shown in FIG. 14, when the number of selections is "2," in addition to the first recommendation level, the corrected second and third recommendation levels are taken into consideration, and the two employees with the highest final recommendation levels, i.e., employee X and employee C, are selected as consultation candidates.

[0121] The consultation candidate selection processing unit 17 also generates consultation history information for each of the selected consultation candidate employees, and stores the generated consultation history information in the consultation history storage unit 34. The consultation history information includes, for example, identification information of the person seeking advice, identification information and attribute information of the consultation candidate, information indicating the problem that the person is consulting, and information indicating whether the consultation was successful or not.

[0122] Information indicating whether the consultation has been successful or not is registered, for example, after the consultation is completed, by sending questionnaire information from the server SV to the user terminal UTi of the person seeking consultation inquiring whether the problem has been solved or not, and the person seeking consultation returns answer information to the questionnaire information, which is then received by the server SV.

[0123] (7) Presentation of Consultation Candidate Information When the consultation candidate is selected, the control unit 1 of the server device SV finally generates presentation information of the selected consultation candidate in step S7 under the control of the selection information output processing unit 18. For example, the presentation information is generated in which the employee information of the selected two consultation candidates is arranged in descending order of recommendation level. Note that the presentation information may also include the recommendation level of the employees selected as consultation candidates.

[0124] The selected information output processing unit 18 transmits the generated presentation information to the user terminal UTi used by the employee who made the consultation from the communication I / F unit 4, for example, by email.

[0125] When the user terminal UTi receives the presented information, it displays the received presented information on its display. At this time, if the presented information includes access information of the candidate employee, the person seeking advice can easily access the user terminal of the candidate employee by clicking on the access information. Note that the person seeking advice can also access the candidate employee by telephone, email, chat, etc.

[0126] (Effect) As described above, in one embodiment, when selecting a candidate employee to consult from among a plurality of other employees in response to a request to select a person to consult about a problem that the employee requesting consultation is having, a pairing process is first used to calculate the frequency of appearance of a group of problem words of the employee requesting consultation in each piece of text data related to the past work of the other plurality of employees, and a candidate employee to consult is selected from among the other plurality of employees based on the calculated frequency of appearance, and a first recommendation level for the selected candidate employee to consult is set based on the degree of coverage or the degree of handling of the group of problem words in the text data of that employee.

[0127] Next, for the candidate employee, the similarity is calculated between a group of problem words that the candidate employee has consulted about in the past as a consultant and a group of problem words of the original employee, and a first recommendation degree correction process is performed for the candidate employees who have had the problem they consulted about solved in the past, correcting the first recommendation degree to a value that increases as the similarity increases.

[0128] Furthermore, the number of times that each of the candidate employees has been selected as a candidate for consultation in the past for problems with a similarity equal to or greater than a threshold is calculated, and a second recommendation degree correction process is performed to correct the first recommendation degree of the candidate employee according to the calculated number of times to a value that decreases as the number of times increases.

[0129] Finally, a certain number of employees are selected as consultation candidates from among the candidate employees in descending order of the second recommendation degree and the third recommendation degree after the correction, and information about the selected employees is presented to the employee who made the consultation.

[0130] Therefore, according to one embodiment, among multiple employees selected as consultation candidates based on the frequency of appearance of a group of problem words, the first recommendation level of an employee who has previously consulted on a problem and had it resolved is corrected based on the similarity of the group of problem words so that the higher the similarity, the higher the first recommendation level. In other words, among multiple employees selected as consultation candidates, an employee who is likely to have consulted on a similar problem in the past and has previously had the problem resolved through that consultation can be selected as a consultation candidate with a higher probability. As a result, the concentration of consultations on specific employees is reduced, thereby reducing the burden on specific employees in responding to consultations and making it possible to prevent a decline in work efficiency.

[0131] Furthermore, for employees who have been selected as candidates for consultation in the past for problems that are highly similar to those of the employee requesting consultation, the first recommendation degree of the employee is corrected according to the number of times the employee has been selected as a candidate for consultation, so that the greater the number of times, the lower the first recommendation degree of the employee. In other words, the recommendation degree of an employee who has received many consultations about similar problems in the past is corrected to be lower. Therefore, the probability that a specific employee will be selected as a candidate for consultation is reduced, which makes it possible to further prevent the concentration of consultation response work on a specific employee.

[0132] [Other embodiments] (1) In one embodiment, a case where both the first recommendation level correction process and the second recommendation level correction process are executed is described, but it is also possible to execute only one of the first recommendation level correction process and the second recommendation level correction process.

[0133] (2) In one embodiment, the highest value among the similarities of the multiple problems is selected as the representative similarity. However, the average value of the similarities of the multiple problems may be calculated and used as the representative similarity.

[0134] (3) In one embodiment, the unit correction amount used in the second recommendation level correction process is set to the same value for all employees, but it may also be set to a different value for each employee. For example, the unit correction amount is set to a larger value for employees with a higher workload, based on employee attribute information such as the employee's years of service, hours of work experience, job title, and overtime hours. In this way, the recommendation level for employees with a high workload can be reduced.

[0135] (4) In one embodiment, the process for selecting consultation candidates according to the present invention has been described as being executed on a server computer on the cloud or the web. However, the present invention is not limited to this, and the process may be executed on a client device used by a user. In other words, the present invention may be realized through communication between clients.

[0136] (5) In one embodiment, the process of selecting potential consultants among employees within a company has been described as an example. However, the present invention can also be applied to other cases, such as when selecting potential consultants among students and teachers at an educational institution such as a school.

[0137] In addition, the functions of the information processing device relating to the process of selecting a candidate consultation destination, and the processing procedures and contents thereof can be modified in various ways without departing from the spirit of the present invention.

[0138] Although the embodiments of the present invention have been described in detail above, the above description is merely an example of the present invention in every respect. It goes without saying that various improvements and modifications can be made without departing from the scope of the present invention. In other words, when implementing the present invention, specific configurations according to the embodiments may be appropriately adopted.

[0139] In short, this invention is not limited to the above-described embodiments, and in the implementation stage, the components can be modified and embodied without departing from the spirit of the invention. Furthermore, various inventions can be formed by appropriately combining multiple components disclosed in the above-described embodiments. For example, some components may be omitted from all the components shown in the embodiments. Furthermore, components from different embodiments may be appropriately combined.

[0140] SV...server device UT1 to UTn...user terminal NW...network 1...control unit 2...program storage unit 3...data storage unit 4...communication I / F unit 5...bus 11...terminal handling data acquisition processing unit 12...pairing condition generation processing unit 13...pairing execution processing unit 14...similarity calculation processing unit 15...first recommendation degree correction processing unit 16...second recommendation degree correction processing unit 17...consultation destination candidate selection processing unit 18...selection information output processing unit 31...terminal handling data storage unit 32...pairing condition storage unit 33...consultation destination candidate list storage unit 34...consultation history storage unit

Claims

1. An information processing device that, in response to a request for selecting a consultation destination for a first problem of a first user, selects consultation candidates from among a plurality of second users and presents them to the first user, comprising: a first candidate selection processing unit that selects a plurality of first candidates to be consultation candidates from among the plurality of second users based on the frequency of appearance of words related to the first user's problem in each text data related to past work of the plurality of second users, and calculates a first recommendation level for each of the selected first candidates based on the coverage or handling level of words related to the first problem in each text data; and a similarity calculation processing unit that calculates a similarity of words related to the first problem of the first user who is the source of the consultation for each of the first candidates to words related to a second problem about which the first candidate has been consulted in the past; an information processing device comprising: a first recommendation degree correction processing unit that corrects the first recommendation degree to a second recommendation degree, the value of which increases as the similarity increases, for a candidate among the first candidates that has resolved the second problem that was previously consulted about; and a second candidate selection processing unit that selects a second candidate from the first candidates based on the second recommendation degree.

2. An information processing device as described in claim 1, further comprising: a selection count calculation processing unit that calculates, for each of the first candidates, the number of times that the first candidate has been selected in the past as the first candidate for the second problem for which the similarity is equal to or greater than a predetermined threshold; and a second recommendation level correction processing unit that corrects the first recommendation level for the first candidate to a third recommendation level that is a lower value the more the number of times is, in accordance with the calculated number of times; wherein the second candidate selection processing unit selects a second candidate from among the first candidates based on the second recommendation level and the third recommendation level.

3. The information processing device of claim 2, further comprising a presentation processing unit that presents the selected second candidate to the first user who is requesting consultation in order of the second recommendation degree and the third recommendation degree calculated for the second candidate.

4. An information processing method for selecting and presenting to a first user, from among a plurality of second users, candidates for consultation in response to a request for selecting a consultation destination for a first problem of a first user, by an information processing device, the method comprising the steps of: selecting a plurality of first candidates as consultation destination candidates from among a plurality of second users based on the frequency of appearance of words related to the first problem of the first user in each text data related to past work of the plurality of second users, and calculating a first recommendation degree for each of the selected first candidates based on the coverage or handling degree of words related to the first problem in each text data; and calculating a similarity of words related to the first problem of the first user, which is the source of consultation, for each of the first candidates, to words related to a second problem about which the first candidate has been consulted in the past; a step of correcting the first recommendation degree to a second recommendation degree, the second recommendation degree being a value that increases as the similarity degree increases, for a candidate among the first candidates that has resolved the second problem previously consulted about; and a step of selecting a second candidate from the first candidates based on the second recommendation degree.

Images