Information processing system, information processing method, and program

The information processing system effectively classifies business names into groups using vector-based AI methods, addressing the challenge of grouping sole proprietors and enhancing operational efficiency.

JP2025142560APending Publication Date: 2025-10-01MONEY FORWARD INC
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2024041997
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2024-03-18
Publication Date
2025-10-01

AI Technical Summary

Technical Problem

Existing systems struggle to effectively group business partners, especially sole proprietors, due to limited publicly available information.

Method used

An information processing system that utilizes a processor to convert names into vectors and classify them into groups based on vector distances, employing AI technology and methods like k-means clustering.

Benefits of technology

Enables accurate classification of names into meaningful groups, allowing for optimized operations based on business characteristics, with the ability to adjust and refine groupings as needed.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2025142560000001_ABST
    Figure 2025142560000001_ABST
Patent Text Reader

Abstract

To provide an information processing system or the like capable of grouping multiple names into multiple groups.SOLUTION: Provided is an information processing system having at least one processor. In the information processing system, the processors input multiple names to be grouped into vectorizing means in an input step. The vectorizing means converts an inputted character string to a vector, and generates a vector in which distance becomes closer as the meaning expressed by the character string is closer. In a grouping step, multiple names are grouped into multiple groups on the basis of distance among multiple vectors generated from the multiple names.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical Field]

[0001] The present invention relates to an information processing system, an information processing method, and a program. [Background technology]

[0002] Patent Document 1 discloses a technique for displaying an organizational chart by grouping customers by department based on customer information obtained from business cards and customer information acquired from the signature portion of emails. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-26405 Summary of the Invention [Problem to be solved by the invention]

[0004] For example, when a business partner is a sole proprietor, it is difficult to group them together because there is little publicly available information about sole proprietors.

[0005] In view of the above circumstances, the present invention provides an information processing system and the like that can classify a plurality of names into a plurality of groups. [Means for solving the problem]

[0006] According to one aspect of the present invention, there is provided an information processing system including at least one processor. In this information processing system, in an input step, the processor inputs a plurality of names to be classified to a vectorization means. The vectorization means converts the input character strings into vectors, and generates vectors whose distances are closer as the meanings expressed by the character strings become closer. In a classification step, the plurality of names are classified into a plurality of groups based on the distances between each of a plurality of vectors generated from the plurality of names.

[0007] According to this aspect, a plurality of names can be classified into a plurality of groups. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a diagram showing an example of the overall configuration of a classification processing system 1. FIG. [Figure 2] 2 is a diagram illustrating an example of a hardware configuration of a server device 10. FIG. [Figure 3] FIG. 2 is a diagram illustrating an example of a hardware configuration of a user terminal 20. [Figure 4] FIG. 10 is an activity diagram illustrating an example of classification processing. [Figure 5] FIG. 10 is a diagram illustrating an example of a system screen. [Figure 6] FIG. 10 is a diagram illustrating an example of a classification result screen. [Figure 7] FIG. 10 is a diagram illustrating an example of a re-sorting screen. [Figure 8] FIG. 10 is a diagram showing another example of the re-sorting screen. [Figure 9] FIG. 10 is a diagram illustrating an example of a system screen. [Figure 10] FIG. 10 is a diagram illustrating an example of a classification table. [Figure 11] FIG. 10 is a diagram illustrating an example of a classification table. [Figure 12] FIG. 10 is a diagram illustrating an example of a processing table. DETAILED DESCRIPTION OF THE INVENTION

[0009] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the accompanying drawings. Various features shown in the following embodiments can be combined with each other.

[0010] Incidentally, the program for realizing the software appearing in one embodiment may be provided as a non-transitory computer-readable medium, or may be provided so that it can be downloaded from an external server, or may be provided so that the program is started on an external computer and its functions are realized on a client terminal (so-called cloud computing).

[0011] Furthermore, various information processing according to an embodiment may realize input and output corresponding to the input. Here, the form of information referenced in such information processing (hereinafter referred to as reference information) is not limited as long as an output is obtained as a result of the input. The reference information may be, for example, rule-based information such as a database, a lookup table, or a predetermined function (including a decision formula such as a regression formula constructed using a statistical method), a trained model that has previously trained the correlation between input and output, or a large-scale language model that can output a desired result by inputting a prompt.

[0012] In one embodiment, a "unit" may include, for example, a combination of hardware resources implemented by a circuit in the broad sense and software information processing that can be specifically realized by these hardware resources. In one embodiment, various information is handled, and this information is represented, for example, by physical values ​​of signal values ​​representing voltage and current, high and low signal values ​​as a binary bit set consisting of 0 or 1, or quantum superposition (so-called quantum bits), and communication and calculations can be performed on a circuit in the broad sense.

[0013] Furthermore, a circuit in the broad sense is a circuit realized by at least an appropriate combination of a circuit, circuitry, processor, memory, etc. The processor may be a general-purpose processor or a dedicated circuit. That is, it includes an application specific integrated circuit (ASIC), a programmable logic device (e.g., a simple programmable logic device (SPLD), a complex programmable logic device (CPLD), and a field programmable gate array (FPGA)), etc.

[0014] <Embodiment> 1. System Configuration The system configuration according to the embodiment will be described below. Fig. 1 is a diagram showing an example of the overall configuration of a classification processing system 1. Fig. 1 shows an overview of each device included in the classification processing system 1 and users who use those devices. Each overview will be explained as needed, with reference to other figures.

[0015] The classification processing system 1 is an information processing system that performs information processing such as grouping multiple names to be classified, i.e., classifying them into groups. The multiple names can include a variety of names, but the following description will use an example in which the names of businesses, including sole proprietors, are classified. The classification processing system 1 also provides a specified service, and is used, for example, by allowing users of the service to group the objects to be classified and then perform operations optimized for each group for the objects (e.g., businesses) represented by the names classified into each group.

[0016] The classification processing system 1 includes a communication line 2, a server device 10, a user terminal 20, and an operator terminal 30. The communication line 2 is not particularly limited, but may be configured, for example, by the Internet network. The communication line 2 may also include a local area network, a mobile communication network, a VPN (Virtual Private Network), etc. The communication line 2 mediates the exchange of data between devices connected to the communication line. In the example of FIG. 1, the server device 10 is connected to the communication line 2 by a wire, and the user terminal 20 and the operator terminal 30 are connected wirelessly. Note that the connection of each device to the communication line 2 may be wired or wireless.

[0017] The server device 10 is an information processing device that executes information processing such as classification processing and service processing for providing the above-mentioned services. The server device 10 stores a name database DB1. The name database DB1 stores a plurality of names (e.g., names of businesses) to be classified.

[0018] The server device 10 also executes information processing using AI (Artificial Intelligence) technology. The server device 10 includes an artificial intelligence module 100. The artificial intelligence module 100 is a module adjusted to realize a predetermined function using AI technology, and hereinafter may be simply referred to as "AI." The artificial intelligence module 100 is adjusted to realize, for example, a vectorization function (described in detail below) that, when a name is input, generates a vector corresponding to the name.

[0019] The user terminal 20 is a terminal whose user is a user of the services provided by the classification processing system 1, and is, for example, a personal computer, a smartphone, or a tablet terminal. The user terminal 20 displays a service usage screen and accepts operations by service users. The operator terminal 30 is a terminal whose user is a system operator who operates the classification processing system 1, and is, for example, a personal computer. The operator terminal 30 displays an operation screen for operating the system and accepts operations by the system operator.

[0020] The server device 10 performs a display process for displaying images on the user terminal 20 and the operator terminal 30, and an authentication process for authenticating the service users and system operators who use the user terminal 20 and the operator terminal 30.

[0021] The server device 10 performs processes such as generating and transmitting an HTML (Hyper Text Markup Language) file as display processing, and uses the browser functions to display a web page showing a system screen on the user terminal 20 and the operator terminal 30. Note that the user terminal 20 and the operator terminal 30 may install an application program for using the classification processing system 1, and the server device 10 may perform processes such as generating and transmitting display data in that application as display processing. The server device 10 controls the display of the user terminal 20 and the operator terminal 30 by performing these display processes.

[0022] The server device 10 stores authentication information (user ID, password, etc.) for authenticating users who use the classification processing system 1, such as service users and system operators, and authenticates users who input the authentication information. By authenticating users, the server device 10 can restrict access to data and assign identification information to data entered by users to make the data identifiable.

[0023] 2. Hardware Configuration The hardware configuration according to the first embodiment will be described below. 2 is a diagram showing an example of the hardware configuration of server device 10. Server device 10 includes a control unit 11, a storage unit 12, a communication unit 13, and a bus 14. Bus 14 electrically connects the various units included in server device 10.

[0024] (Control unit 11) The control unit 11 has at least one processor. The at least one processor may be configured by, for example, a central processing unit (CPU), a micro processing unit (MPU), a graphics processing unit (GPU), one or more integrated circuits, one or more discrete circuits, or a combination thereof (not shown).

[0025] The control unit 11 is a computer that realizes various functions related to the classification processing system 1 by reading out predetermined programs stored in the storage unit 12. In other words, information processing by software stored in the storage unit 12 is specifically realized by the control unit 11, which is an example of hardware, and can be executed as each functional unit included in the control unit 11. Note that the control unit 11 is not limited to being a single unit, and it may be implemented with multiple control units 11 for each function. It may also be a combination of these.

[0026] (Storage unit 12) The memory unit 12 stores various pieces of information defined above. This can be implemented, for example, as a storage device such as a solid state drive (SSD) or a hard disk drive (HDD) that stores various programs and the like related to the classification processing system 1 executed by the control unit 11, or as a memory such as a random access memory (RAM) that stores temporarily required information (arguments, arrays, etc.) related to the program operations. The memory unit 12 stores various programs, variables, etc. related to the classification processing system 1 executed by the control unit 11.

[0027] (Communications Department 13) The communication unit 13 is configured by a communication module. The communication module may be a wireless communication module conforming to standards such as IEEE802.11a / b / g / n / ac / ax, LTE, 5G, or 6G, or may be a wired communication module conforming to standards such as IEEE802.3. The communication unit 13 is configured to be able to transmit various electrical signals from the server device 10 to external components. The communication unit 13 is also configured to be able to receive various electrical signals from the external components to the server device 10. More preferably, the communication unit 13 has a network communication function, which allows various information to be communicated between the server device 10 and external devices via the communication line 2.

[0028] Fig. 3 is a diagram showing an example of the hardware configuration of user terminal 20. User terminal 20 includes control unit 21, memory unit 22, communication unit 23, input unit 24, output unit 25, and bus 26. Bus 26 electrically connects the various units included in user terminal 20. Control unit 21, memory unit 22, and communication unit 23 are similar hardware to control unit 11, memory unit 12, and communication unit 13 shown in Fig. 2, although their specifications, models, etc. may differ.

[0029] (Input unit 24) The input unit 24 has keys, buttons, a touch screen, a mouse, etc., and receives input from the user. The input unit 24 may also have a microphone and have the function of receiving voice input from the user.

[0030] (Output section 25) The output unit 25 has a display, a speaker, etc., and displays visual information generated in a manner that is visible to the user, such as a screen, an image, an icon, or text, on the display surface of the display, and outputs sounds including voice.

[0031] 3 has the same hardware configuration as the user terminal 20. In the following description of the operator terminal 30, only the control unit 31 is assigned a different reference numeral from the control unit 21 of the user terminal 20.

[0032] 3. Information Processing The information processing according to the embodiment will be described below. In the following description, the server device 10, the user terminal 20, and the operator terminal 30 are described as the subjects of each information processing, but the information processing is executed by at least one processor included in the classification processing system 1, i.e., the processor included in the control unit of each device. The classification processing system 1 executes the classification processing described above.

[0033] 4 is an activity diagram showing an example of classification processing. The classification processing is executed when the system operator performs a grouping operation on the operator terminal 30. The grouping operation is performed while a system screen for grouping is displayed on the operator terminal 30.

[0034] Fig. 5 is a diagram showing an example of a system screen. The system screen C1 shown in Fig. 5 displays a character string "Please enter classification information," an input field D11 for classification targets, a display field D12 for a list of names, a display field D13 for the number of names, an input field D14 for the number of groups, and an execute grouping button B11. The "classification information" referred to here refers to information necessary for performing classification processing, such as the classification targets and the number of groups.

[0035] The input field D11 is a field for inputting information for specifying names to be categorized. There are various methods for specifying names, but for example, in the case of business names, the name of the company to which the system operator belongs (businesses that do business with this company are specified), the region, size, etc. are input as classification information. Note that if all names stored in the name database DB1 are to be categorized, "all names" can be input as classification information. The operator terminal 30 accepts the specification of names to be categorized based on the operation of inputting this classification information (activity A11). The operator terminal 30 transmits the input classification information to the server device 10.

[0036] The server device 10 reads the names indicated by the transmitted classification information from the name database DB1 and transmits them to the operator terminal 30 (activity A12). The operator terminal 30 displays the transmitted list of names and the number of names as name information in display fields D12 and D13, respectively (activity A13). The system operator looks at the displayed list of names and the number of names and determines whether they are appropriate for classification.

[0037] When the system operator determines that the classification targets are appropriate, he / she inputs the number of groups he / she wishes to specify in input field D14 as classification information. The number of groups to be specified may be determined based on the number of industries to be classified, or may be determined based on the approximate number of businesses per group. In either case, the number of groups can be changed later, so a provisional number of groups should be input first. The operator terminal 30 accepts the number of groups specified based on this input operation by the system operator (activity A14).

[0038] After specifying the number of groups, the system operator operates the execute button B11 to execute the classification process. When the operator terminal 30 receives the grouping instruction, it transmits instruction data to the server device 10 to instruct it to classify the specified classification targets (plurality of names) into the specified number of groups (activity A15). The instruction data is, for example, data indicating all of the multiple names, but it may also be information capable of identifying the multiple names stored in the name database DB1 (for example, information indicating classification information).

[0039] When the server device 10 receives the transmitted instruction data, it inputs the classification target indicated by the received instruction data to the artificial intelligence module 100 (activity A21). Next, the server device 10 instructs the artificial intelligence module 100 to vectorize the input classification target (activity A22). Upon receiving this instruction, the artificial intelligence module 100 executes vectorization processing on the input classification target (activity A23).

[0040] As described above, the artificial intelligence module 100 has a vectorization function that vectorizes names. This vectorization function is realized, for example, by an embedding function, which is one of the functions of a natural language processing model whose accuracy has been improved by machine learning using a large data set called LLM (Large Language Models). The embedding function converts complex data such as words and phrases into an easily processable numerical vector representation. With the embedding function, the closer the meanings expressed by the character strings, the closer the distance between the generated vectors. Note that vectorization using the embedding function is just one example, and vectorization may also be performed using other well-known techniques.

[0041] The artificial intelligence module 100 outputs vectors converted from each name generated by the vectorization process. The server device 10 executes a process of grouping the multiple names to be classified based on the output vectors (activity A24). The server device 10 executes the grouping process using, for example, the k-means method.

[0042] In this case, the server device 10 first randomly allocates clusters to the objects to be classified. The number of allocated clusters at this time is the same as the specified number of groups. Next, the server device 10 calculates the center of gravity of the data allocated to each cluster. Subsequently, the server device 10 reallocates each data to the cluster with the closest center of gravity. The server device 10 then repeats the calculation of the cluster center of gravity and the reallocation of clusters until the clusters of all data remain unchanged. In this way, the server device 10 classifies the multiple names to be classified into the specified number of groups.

[0043] The vector-based classification method is not limited to the k-means method. For example, well-known classification methods such as the k-means++ method, the k-nearest neighbor method, hierarchical clustering, DBSCAN (Density-Based Spatial Clustering of Applications with Noise), MeanShift, or Gaussian mixture model may be used. A rule-based classification method may also be used, utilizing the rules of the multiple names to be classified. A rule-based classification method is, for example, a method in which combinations of words that may be included in names are associated with groups, and names are classified into groups based on the words included in the names to be classified.

[0044] After completing the grouping, the server device 10 next extracts label candidates from the names belonging to each classified group (activity A25). A label is, for example, a word or phrase that indicates what kind of name belongs to a group. The server device 10 extracts, for example, a word or phrase that is commonly used among names classified into the same group as label candidates. For example, since "restaurant," "dining hall," and "cafe" are extracted from names classified into the "food service industry," the server device 10 extracts, for example, the most frequently used word or phrase among them as label candidates.

[0045] Note that the method of extracting label candidates is not limited to this. For example, the server device 10 may extract, as a label candidate, a word or phrase having the smallest statistical value, such as the average value or minimum value of the distance from the average vector, among the extracted words or phrases. Next, the server device 10 generates screen data showing a screen of the classification results using the results of the grouping process and the extracted label candidates (activity A31). The server device 10 transmits the generated screen data to the operator terminal 30. The operator terminal 30 displays the classification result screen shown in the transmitted screen data (activity A32).

[0046] Fig. 6 is a diagram showing an example of a classification result screen. The classification result screen C2 shown in Fig. 6 displays a character string "Individual classification results," a group name display field D21, a name list display field D22, a name count display field D23, a name variance index display field D24, a label display field D25 for each group, a re-classification check field D26, a classification number display field D27, a classification result save button B21, a return to classification target specification operation button B22, an individual re-classification button B23, and a full re-classification button B24.

[0047] Display field D21 displays the identification information of each classified group (such as "Group 1" in the example of FIG. 6). Display field D22 displays a list of the names classified into each group. The list of groups and names displayed in FIG. 6 is only a partial list, and the remaining groups and names can be displayed by scrolling. Display field D23 displays the number of names classified into each group.

[0048] The display field D24 displays a dispersion index, which is an index indicating the degree of dispersion of vectors converted from names classified into each group. The dispersion index is expressed, for example, by the average value of the distance between each vector and the average vector of all name vectors belonging to each group. The larger this distance, the more dispersed the vectors converted from names in the same group. The distance between vectors is expressed, for example, by Euclidean distance. However, without being limited to this, the distance between vectors may also be expressed by other calculation methods, such as Manhattan distance or Chebyshev distance.

[0049] Furthermore, the method for calculating the dispersion index is not limited to the above method. For example, a statistical value such as the maximum or median of the distances between the average vector and each vector may be calculated as the dispersion index. Alternatively, a statistical value such as the average, maximum, or median of the distances for all combinations of two vectors that can be extracted from all vectors belonging to the group may be calculated as the dispersion index. In short, any value that indicates an index of the degree of dispersion of vectors within a group may be calculated as the dispersion index.

[0050] Display field D25 displays the label candidates extracted in A25. The display in display field D25 can be edited by the system operator. The system operator can look at the list of names displayed in display field D22 to determine whether the label candidates are appropriate and make corrections if necessary. Check field D26 displays a check image indicating a group that requires reclassification. If the system operator looks at the number of names and the dispersion index and determines that the group is not appropriate, he or she can check check field D26 to redo the classification of the names belonging to the checked group (i.e., reclassify them).

[0051] By operating the back button B22, the user can return to the system screen C1 and re-enter the classification information. The operator terminal 30 determines whether or not the operation on the back button B22, i.e., the operation to re-designate the classification target, has been accepted (activity A33), and if it determines that the operation has been accepted (YES), it returns to A11 (designation of classification target) or A14 (designation of the number of groups) and performs the operation. If the operator terminal 30 determines that the operation to re-designate has not been accepted (NO), it next determines whether or not the operation to re-classify has been accepted (activity A34).

[0052] If reclassification is not necessary, the system operator does not perform the reclassification operation (A34=NO), but operates the save button B21 to save the classification results. When the operator terminal 30 accepts the save operation (activity A51), it transmits request data to the server device 10 requesting that the classification results be saved in association with the classification number (identification information for identifying the classification results) displayed in the display field D27. The server device 10 saves the classification results in association with the classification number indicated in the transmitted request data (activity A52).

[0053] For example, if the system operator determines that a group with a particularly large dispersion index needs to be reclassified, the system operator operates the individual reclassification button B23. When the operator terminal 30 determines that the reclassification button B23 has been operated, i.e., that a reclassification operation has been performed (A34=YES), the system operator transmits to the server device 10 individual classification request data requesting reclassification of the names belonging to the group for which the check box D26 is checked. Upon receiving the individual classification request data, the server device 10 executes individual reclassification processing for the requested group (activity A40). The server device 10 first displays, on the operator terminal 30, a reclassification screen, which is a screen for performing individual reclassification.

[0054] Fig. 7 is a diagram showing an example of a re-categorization screen. The re-categorization screen C3 shown in Fig. 7 displays a character string "Please select a re-categorization method for each group," a display field D31 for the group to be re-categorized, a display field D32 for the name etc. to be re-categorized, a group switching button B31, a button B32 for returning to the classification result screen, an execute button B33 for re-categorizing according to the system operator's selection, and an execute button B34 for re-categorizing all automatically.

[0055] Display field D31 displays one of the groups whose check box D26 in FIG. 6 was checked. The group displayed in display field D31 can be switched by operating switch button B31. Display field D32 displays, from among the names belonging to the currently displayed group, names whose degree of deviation from the group of names belonging to the group (deviance) is equal to or greater than a predetermined degree. The degree of deviation is indicated, for example, by the distance from the average vector of all names belonging to the group, and the greater this distance, the greater the degree of deviation.

[0056] For each name, you can select "Manual Reclassification," "Automatic Reclassification," or "Maintain." In the "Manual Reclassification" column, a certain number of groups (three in the example of Figure 7) other than the group to which each name is currently classified, which have a vector distance close to the name, are displayed as candidates for reclassification. The distance between vectors here refers to, for example, the distance between the name's vector and the average vector of the group. For example, the name "AAAA" is currently classified as "Retail," but "Manufacturing" and "Food Service" are displayed as candidates for reclassification.

[0057] Additionally, the "Automatic Reclassification" column displays the group other than the group to which each name is currently classified, that has the closest distance between the vector and the name with the highest degree of outlier. For example, for the name "AAAA," "Manufacturing" is displayed as the group to which it will be automatically reclassified. Additionally, the "Maintain" column displays the name of the group to which each name is currently classified ("Retail" in the example in Figure 7). The "Manual Reclassification," "Automatic Reclassification," and "Maintain" display columns have thicker borders to make it clear which parts have been selected by the system operator.

[0058] When the execute button B33 is operated in this state, reclassification is performed for each name using a method selected by the system operator, or the current classification results are maintained without reclassification. Specifically, the operator terminal 30 accepts operations on the display field D32 and the execute button B33 as reclassification operations, and transmits request data requesting reclassification using the selected content to the server device 10 (activity A34). The server device 10 executes the reclassification process indicated by the transmitted request data (activity A40).

[0059] The server device 10 reclassifies names for which reclassification has been manually selected in the display field D32 into the selected reclassification candidate. Furthermore, the server device 10 reclassifies names for which reclassification has been automatically selected in the display field D32 into the group with the next shortest vector distance from the current group. Furthermore, the server device 10 does not reclassify names for which maintenance has been selected in the display field D32, but maintains the current classification.

[0060] When the execute button B34 in Figure 7 is operated, the server device 10 automatically reclassifies all names that belong to the displayed group and have a high degree of deviation, that is, reclassifies them into the group with the next shortest vector distance from the current group. After the system operator has performed reclassification as needed for the group currently being displayed, he or she operates the switch button B31 to perform reclassification for other groups as needed. When the back button B32 is operated, the server device 10 ends the reclassification process and returns to A31 to generate screen data of the classification results. The classification results in this case will be the classification results after reclassification.

[0061] On the classification result screen C2 shown in Fig. 6, for example, if the system operator determines that the variance index is large overall and that reclassification is necessary for all groups, the system operator operates the overall reclassification button B24. When the operator terminal 30 determines that the reclassification button B24 has been operated, i.e., that a reclassification operation has been performed (A34 = YES), the system operator transmits overall classification request data to the server device 10, requesting reclassification of the names belonging to all groups. Upon receiving the overall classification request data, the server device 10 executes the requested overall reclassification process (activity A40). The server device 10 first displays a reclassification screen, which is a screen for performing overall reclassification, on the operator terminal 30.

[0062] Fig. 8 is a diagram showing another example of the re-classification screen. The re-classification screen C4 shown in Fig. 8 displays a character string saying "Please enter the number of groups when re-classifying all," a display field D41 showing a summary of the classification results, a display field D42 showing a graph of the classification results, an input field D43 for the new number of groups, a back button B41 for returning to the classification result screen shown in Fig. 6, and an execute re-classification of all button B42.

[0063] Display field D41 displays the "classification number," "number of names," and "number of groups," as well as the "average number of names," "maximum number of names," and "minimum number of names" across all groups, and the "average variance index," "maximum variance index," and "minimum variance index" across all groups. Display field D42 also displays a graph showing the number of groups within a common range of the number of names belonging to the groups, and the number of groups within a common range of the variance index of the groups. Using this information as a reference, the system operator enters a new number of groups into input field D43 and operates execute button B42.

[0064] The operator terminal 30 accepts these operations as reclassification operations and transmits request data to the server device 10 requesting reclassification using the input new number of groups (activity A34). The server device 10 reclassifies using the number of groups indicated in the transmitted request data. In this case, the server device 10 may display a summary and graph based on the reclassified groups while keeping the reclassification screen C4 displayed, or may display the classification result screen C2 shown in FIG. 6.

[0065] As described above, the server device 10 functions as an example of an input unit that inputs multiple names to be classified to vectorization means. The artificial intelligence module 100 has the embedding function described above and is an example of vectorization means that converts input character strings into vectors. As described above, this vectorization means is a means for generating vectors whose distances become closer as the meanings expressed by character strings become closer. For example, at A21 shown in FIG. 4, the server device 10 inputs multiple names to be classified to the artificial intelligence module 100.

[0066] Furthermore, the server device 10 functions as an example of a classification unit that classifies a plurality of names into a plurality of groups based on the distances between each of a plurality of vectors generated from the plurality of names. The server device 10 groups a plurality of names to be classified by the grouping process A24 shown in Fig. 4. The grouping results are displayed in a group name display column D21 and a name list display column D22 shown in Fig. 6. According to this embodiment, a plurality of names can be classified into a plurality of groups as shown in Fig. 6.

[0067] Furthermore, the server device 10 (an example of a classifier) ​​performs classification using a method (e.g., K-means) for classifying into a specified number of groups. Then, the server device 10 (an example of a classifier) ​​reclassifies by changing the number of groups. In the example of FIG. 8, the server device 10 reclassifies to the number of groups entered in the input field D43. According to this aspect, the number of groups can be made more appropriate compared to when reclassification by changing the number of groups is not performed.

[0068] Furthermore, the server device 10 (an example of a classification unit) reclassifies a name converted into a vector whose distance to other vectors in the same group is equal to or greater than a first threshold into a group among other groups in which the distance between the name and the vector is the closest. For example, when the execute button B34 (a button for automatically reclassifying all) shown in FIG. 7 is operated, the server device 10 reclassifies the name displayed in the display field D32 into the group whose distance between the name and the vector is the closest (for example, for a name "AAAA," into the group "Manufacturing"). The display field D32 displays names whose "outlier degree," which represents the distance from other vectors in the same group, is equal to or greater than a predetermined degree (an example of a first threshold). According to this embodiment, it is possible to reduce the number of names that are significantly outliers within a group compared to when reclassification is not performed.

[0069] Furthermore, the server device 10 functions as an example of a first display unit that displays a name converted into a vector whose distance from other vectors in the same group is equal to or greater than a second threshold, and one or more other groups (the re-categorization candidates shown in FIG. 7 are an example of this) from among the other groups whose distance from the name to the vector is closest. The re-categorization candidates shown in FIG. 7 represent a name with a high degree of outlier and one or more other groups whose distance from the vector is closest (for example, for the name "AAAA," "manufacturing industry" and "food service industry"). In the display field D32, names whose "outlier degree," which represents the distance from other vectors in the same group, is equal to or greater than a predetermined degree (an example of the second threshold) are displayed as re-categorization candidates.

[0070] Then, the server device 10 (an example of a classification unit) reclassifies the converted name into a group selected by the user from one or more other groups that have been displayed. In the example of Fig. 7, the server device 10 reclassifies the name "AAAA" into the group "food service industry" selected by the system operator. According to this embodiment, the effort required by the user to reclassify the name can be reduced compared to when reclassification candidates are not displayed.

[0071] The server device 10 also functions as an example of a first extraction unit that extracts one or more words commonly used in names classified into the same group. The server device 10 also functions as an example of a second display unit that displays the extracted one or more words as label candidates for the same group. The server device 10 extracts words from the names in the group in A25 shown in FIG. 4 and displays the extracted words as label candidates in a display field D25 shown in FIG. 6. Note that, although the server device 10 displays only one label candidate in the example of FIG. 6, it may also display multiple label candidates and allow the system operator to select one. This aspect can assist in determining a label.

[0072] <Variation: Regarding the name> Names to be classified are not limited to names of businesses. Names to be classified may be, for example, store names, organization names, product names, food names, and chemical compound names. For example, business names are often named to allow users to understand the nature of their business, and so they contain words or phrases that describe the nature of their business, making it possible to group them based on their names. In addition, other names also have certain rules for naming, and tend to contain the same words or words with similar meanings in the same group. Any name that has such characteristics may be used as a classification target.

[0073] <Variation: Regarding reclassification> The reclassification may be performed in a manner other than the above. For example, the server device 10 (an example of a classification unit) calculates an index value for each group that indicates the magnitude of the distance between vectors in the same group. For example, the dispersion index (an index indicating the degree of dispersion of vectors) shown in FIG. 6 is an example of the index value. In the example of FIG. 6 etc., the server device 10 displays the calculated dispersion index as an index value, making it easier for the system operator to select targets for reclassification.

[0074] Without being limited to this, for example, for a group whose calculated dispersion index is larger than a predetermined standard, the server device 10 may encourage reclassification by checking and displaying a reclassification button B24 shown in Fig. 6 in advance. The predetermined standard may be determined, for example, by repeatedly classifying sample names and verifying the dispersion index value at which inappropriate classification is more likely to occur.

[0075] Furthermore, the server device 10 (an example of a classification unit) may reclassify when the calculated index value is greater than a predetermined standard. For example, the server device 10 determines that the index value is greater than the predetermined standard when the dispersion index of one of all groups or a predetermined percentage or more of the groups is greater than a threshold. For example, when the calculated index value is greater than the predetermined standard, the server device 10 may automatically reclassify (either individual reclassification or overall reclassification) without any operation by the system operator. When performing individual reclassification, the server device 10 may, for example, reclassify groups whose index values ​​are greater than the predetermined standard.

[0076] Furthermore, before automatically reclassifying, the server device 10 may display a confirmation image to ask the system operator whether it is OK to reclassify, and may perform the reclassification once confirmation is received. In this case, the server device 10 may display a selection image that allows the user to select whether to perform individual reclassification or overall reclassification (in the case of overall reclassification, an input field for the number of new groups is also provided), and perform the reclassification using the selected method. As described above, by reclassifying classification results with index values ​​greater than a predetermined standard, it is possible to reduce groups with low accuracy.

[0077] The server device 10 may vary the predetermined criterion depending on the number of classified groups. For example, the server device 10 sets a smaller criterion because the greater the number of classified groups, the fewer the number of names belonging to each group and the shorter the distance between vectors. Note that the distance between vectors varies depending on the vectorization means and the distance calculation method, and also varies depending on the tendency of the multiple names to be classified, so the greater the number of groups, the greater the criterion may be.

[0078] Furthermore, the server device 10 may automatically determine the number of new groups when performing overall reclassification as in the example of Fig. 8. The server device 10 determines the number of new groups according to, for example, the magnitude of the dispersion index of the entire group. For example, the server device 10 performs reclassification by increasing the number of new groups as the dispersion index of the entire group increases.

[0079] Furthermore, the server device 10 may determine the number of new groups according to the distance between the vectors of the groups. The distance between the vectors of the groups is represented, for example, by the distance between the average vectors of each group. However, without being limited to this, a statistical value (such as the average or median) of the distance between the vectors for all combinations of the vectors of one group and the vectors of the other group may be used as the distance between the vectors of the groups. In either case, the server device 10 performs reclassification by increasing the number of new groups, for example, as the distance between the vectors of the groups increases.

[0080] As mentioned above, the greater the number of groups, the shorter the distance between vectors and the smaller the dispersion index, so by determining the number of groups as described above, it is possible to make it easier for the dispersion index of each reclassified group to be small. Note that the distance between vectors and the dispersion index also change depending on the vectorization means and distance calculation method, and also change depending on the tendency of the multiple names to be classified, so it may be possible to increase the number of groups as the dispersion index or the distance between vectors of groups becomes smaller.

[0081] The first and second thresholds used to determine whether to reclassify may be set to values ​​according to the number of groups. For example, the server device 10 sets the first and second thresholds smaller as the number of groups increases, so that even if a name has the same distance from other vectors in the same group, the larger the number of groups, the more likely it is to be reclassified.

[0082] In either case, it is desirable to determine the predetermined criterion, the number of groups, and the first or second threshold value so that the variation in the reclassified groups is likely to be small and the likelihood of names belonging to each group being outliers is likely to be small.

[0083] <Variation: 2-stage classification> In the above example, the server device 10 classifies multiple names at once, but this is not limited to this. For example, the server device 10 (an example of a classifier) ​​may classify multiple names into a first number of groups using a method for classifying into a specified number of groups (e.g., the k-means method), and then further classify each group after classification into a second number of groups. A screen displayed in this case will be described with reference to FIG. 9.

[0084] FIG. 9 is a diagram showing an example of a system screen. The system screen C5 shown in FIG. 9 displays a first number input field D54 and a second number input field D55 instead of the group number input field D14 of the system screen C1 shown in FIG. 1. Fixed numbers, for example, are input into the input fields D54 and D55. For example, if "4" is input as the first number and "5" is input as the second number, the server device 10 first classifies the multiple names into four groups, which is the first number, and then further classifies each of the four groups into five groups, which is the second number. As a result, the multiple names are classified into a total of 20 groups.

[0085] The results may differ between classifying into all groups at once (i.e., classifying in one stage) and classifying in two stages as described above. Depending on the tendencies of the multiple names, it may be possible to classify into more appropriate groups by classifying in two stages rather than by classifying in one stage. Therefore, according to this embodiment, depending on the tendencies of the multiple names, it is possible to classify into more appropriate groups than by classifying in detail all at once.

[0086] The two-stage classification method is not limited to the above method. For example, in the example of Fig. 9, the server device 10 requires the user to input the first number and the second number in advance. However, the server device 10 may require the user to input only the first number first, perform classification processing, display the results, and then require the user to input the second number. According to this embodiment, after viewing the first-stage classification results, the user can input a more appropriate number as the second number.

[0087] Further, the server device 10 may vary the second number, which is the number of groups to be classified next, for each group of the first number classified first. For example, the server device 10 (classification unit) calculates an index value that is an index of the magnitude of the distance between vectors in the same group for each of the groups of the first number, and sets the number corresponding to the calculated index value as the second number in each group. In this case, the server device 10 uses a classification table in which the dispersion index and the second number are associated with each other.

[0088] FIG. 10 is a diagram showing an example of a classification table. In the classification table TB1 shown in FIG. 10, classification indexes of "less than TH11", "TH11 or more and less than Th12", and "Th12 or more" are associated with second numbers of "N11", "N12", and "N13" (N is a natural number such that N11 < N12 < N13). The server device 10 calculates the dispersion index of each group, and performs the second-stage classification process using the number associated with the calculated dispersion index as the second number. According to such an aspect, since a group with a larger dispersion index is classified into more groups, the degree of variation within the group can be made uniform compared to the case where the second number is made uniform.

[0089] Further, the server device 10 (classification unit) may count the number of names classified into each group of the groups of the first number, and set the number corresponding to the counted number of names as the second number in each group. In that case, the server device 10 uses a classification table in which the number of names within the group and the second number are associated with each other.

[0090] FIG. 11 is a diagram showing an example of a classification table. In the classification table TB2 shown in FIG. 11, classification indicators such as "less than TH21", "greater than or equal to Th21 and less than Th22", and "greater than or equal to Th22" are associated with a second number such as "N21", "N22", and "N23" (N is a natural number where N21 < N22 < N23). The server device 10 counts the number of names in each group and performs second-stage classification processing using the number associated with the counted number of names as the second number. According to such an aspect, since groups with a larger number of names classified in the first stage are classified into more groups, the scale of each group can be made uniform compared to the case of making the second number uniform.

[0091] <Modified Example: Input of Parameters> In the above example, the plurality of names were each the name of an operator. In that case, in addition to the name of the operator, vectorization may also be performed according to transactions between operators. The name database DB1 stores, in addition to the name of the operator, transaction information of each operator. The transaction information includes, for example, transaction content (such as goods or services traded), transaction partners, and the industry type of the transaction partners.

[0092] The server device 10 functions as an example of an input unit that inputs, in addition to a plurality of names, transaction information regarding transactions conducted by the operators of the plurality of names to the vectorization means. The above-described vectorization means, that is, the Embedding function, generates a vector in which the closer the meaning represented by the character string, the closer the distance. Therefore, for example, the vectors of operators handling the same goods, operators conducting transactions with the same partner, and operators conducting transactions with partners of the same industry type are closer in distance compared to other operators.

[0093] Then, the server device 10 (an example of a classification unit) classifies the plurality of names into a plurality of groups based on the distance between each of the plurality of vectors generated from the plurality of names and transaction information. According to such an aspect, since classification is performed taking into account the actual state of transactions, the accuracy of the groups can be improved compared to the case where transaction information is not input.

[0094] <Variation: Fine Tuning> In the classification processing system 1, an LLM with an embedding function, i.e., an artificial intelligence module 100 that has performed machine learning, is used as vectorization means. The server device 10 may function as an example of a learning unit that causes the artificial intelligence module 100 to learn words related to the field to which the multiple names to be classified belong.

[0095] If the multiple names are, for example, names of businesses, the server device 10 causes the artificial intelligence module 100 to additionally learn technical terms used in each field. For example, in the field of "food service industry," additional terms such as "cafe," "restaurant," "bistro," "dining," and "grill" are additionally learned. In addition, in the field of "medical industry," additional terms such as "clinic," "medical," "health," "care," and "wellness" are additionally learned.

[0096] The server device 10 also performs additional learning on other fields such as retail, beauty, food and beverage, manufacturing, transportation, and accommodation. By additionally learning terms related to each field, the artificial intelligence module 100 can more easily correctly determine the field to which a name containing those terms belongs, compared to when additional learning is not performed, and the distance between vectors representing names of the same field can be shortened, i.e., the accuracy of vectorization can be improved.

[0097] <Variation: Classification into existing labels> In the above example, label candidates were extracted from the words contained in the name, but names can also be classified against existing labels. For example, when the name of a business is the target of classification, the names of each industry defined as a medium classification in the Japan Standard Industrial Classification (accommodation, restaurants, take-out and delivery food services, laundry, hairdressing, beauty and bath services, entertainment, school education, other education and learning support services, medical services, health insurance, etc.) can be used as existing labels.

[0098] First, the server device 10 functions as an example of a generation unit that generates a representative vector representing vectors generated from a plurality of names, based on vectors generated from names that conform to a predetermined rule. The rule here is a rule that is satisfied by names that conform to existing labels. For example, if the medium categories of the Japan Standard Industrial Classification are used as existing labels, the rule is that the names must contain sub-categories belonging to each medium category (for example, if the medium category is "restaurants," then "dining hall," "restaurant," "ryotei," "Chinese restaurant," "ramen restaurant," "yakiniku restaurant," "soba / udon restaurant," "sushi restaurant," "bar," "beer hall," etc.).

[0099] For example, the server device 10 generates a representative vector by using the average vector of the vectors generated from names that conform to the above rules. Then, the server device 10 (an example of a classification unit) classifies names that do not conform to the rules according to the distance between the vector of the name and the representative vector. For example, in the case of names such as "ABC Sushi," "DEF Ramen," and "Grilled Meat GHI," the server device 10 classifies these as "restaurants" because they do not include any of the above minor categories in their names but are closest in distance to the representative vector of "restaurants" among their representative vectors.

[0100] The existing labels are not limited to the above examples, and may be, for example, unique industry classifications provided by private companies. Furthermore, if the classification target is "organization name," the organization name used by a model group of companies may be used as the existing label. Furthermore, if the classification target is "product name," the product name used on an e-commerce site that handles a variety of products may be used as the existing label. According to this embodiment, it is possible to classify items into groups corresponding to existing labels.

[0101] <Variation: Addition of name> The server device 10 may classify the names newly added to the classified groups. In that case, the server device 10, for example, causes the artificial intelligence module 100 to generate vectors of the added names, calculates the distances between the generated vectors and the average vectors of each group, and classifies them into the group with the closest calculated distance. Note that the classification method for new names is not limited to this. The server device 10, for example, may calculate the average distance between the vectors generated for the new names and the vectors of the other names belonging to each group, and classify them into the group with the closest calculated distance.

[0102] <Variation: Post-processing> After classifying a plurality of names into a plurality of groups, post-processing for operations using the classification results may be performed. For example, assume that each of the plurality of names is the name of a business operator, and a predetermined process is associated with each of the plurality of groups. The predetermined process is, for example, a credit process. The credit process is a process for minimizing the risk of transactions with business operators, and includes, for example, a process for obtaining the transaction history of business operators, a process for setting the credit limit of business operators, a process for setting credit conditions, a process for monitoring the financial status, and the like.

[0103] In that case, the server device 10 functions as an example of a processing unit that executes the process associated with the classified group for the business operator of the name classified into that group. The server device 10 uses a processing table that associates the label of the group with the process to be executed.

[0104] FIG. 12 is a diagram showing an example of a processing table. In the classification table TB3 shown in FIG. 12, groups such as "stable group", "normal group", and "volatile group" are associated with processes such as "transaction history acquisition process for the past N31 years,...", "transaction history acquisition process for the past N32 years,...", and "transaction history acquisition process for the past N33 years,..." (N is a natural number such that N31 < N32 < N33).

[0105] The stable group is a group of industries with stable demand and likely stable business performance, while the fluctuating group is a group of industries with large fluctuations in demand and likely unstable business performance. The normal group is a group of industries that fall between the stable group and the fluctuating group. The labels that belong to each group are assumed to be predetermined. The server device 10 executes the processing associated with each group in the classification table TB3 for businesses whose names belong to each group classified by the classification processing.

[0106] In the example of Figure 12, credit processing is carried out based on a short transaction history for businesses with names classified in the stable group, and credit processing is carried out based on a longer transaction history for businesses with names classified in the volatile group. By differentiating the content of credit processing depending on the group, credit operations can be carried out in an appropriate manner for each group of businesses, compared to when a uniform credit processing is carried out, and transaction risks can be reduced.

[0107] 12, the processing to be executed is changed by changing the parameters (the period of the transaction history) used in the transaction history acquisition processing, but the content of the processing itself may also be changed. For example, the transaction history acquisition processing itself may not be performed for the stable group, and the financial data acquisition processing may also be performed for the variable group.

[0108] Furthermore, the processing for business operations according to the classification results is not limited to credit processing. For example, sales email processing, which involves sending emails to customers (including potential customers), may be performed. In this case, the names of those customers are classified into multiple names, and the products sent in the sales emails can be varied depending on the classification group. Any other business processing may be performed as long as it is appropriate to perform different processing depending on the classification group. In either case, different processing can be performed for each group, which makes it easier to achieve business objectives compared to when uniform processing is performed.

[0109] <Variation: Feedback of post-processing results> When the above post-processing is performed, the results of the processing may be fed back. In this case, for example, the server device 10 functions as an example of an evaluation unit that evaluates the results of the processing. For example, when the above credit processing is performed, the server device 10 calculates or obtains the collection period or accounts receivable turnover rate of each group and evaluates them as the results of the credit processing.

[0110] Then, if the evaluation does not meet a predetermined result, the server device 10 (classification unit) reclassifies the groups. For example, the server device 10 executes credit processing for groups classified by the previous classification process, and the evaluation (collection period or accounts receivable turnover rate) of each group is below a predetermined standard. In this case, the server device 10 changes the number of groups from the previous time and executes the classification process. Note that the server device 10 may reclassify only the groups whose evaluations are below the standard, or may display the groups whose evaluations are below the standard and perform the reclassification using a method selected by the system operator. According to such an embodiment, the processing results can be improved.

[0111] <Example of variation: Variation of composition> The configurations (overall configuration, hardware configuration, functional configuration, etc.) shown in FIG. 1 and elsewhere are merely examples, and other configurations may be used as long as they are not inconvenient for implementation. For example, the server device 10, the user terminal 20, and the operator terminal 30 may each be distributed across two or more devices (e.g., the artificial intelligence module 100 may be installed in another device), or may be provided in the form of SaaS (Software as a Service) or a cloud computing system. Furthermore, the information processing performed by the server device 10, the user terminal 20, and the operator terminal 30 may be collectively performed by a device that integrates these devices (a device that integrates two or three of them). In short, as long as the necessary information processing is performed by the classification processing system 1 as a whole, the devices that perform this information processing may have any configuration.

[0112] The output destination of information or data (hereinafter referred to as "information, etc.") may be another device, a display, a memory unit (including an internal memory unit and an external memory unit), an email address, an account of another system, etc. Acquisition of information, etc. includes acquiring information, etc. generated by the device itself, as well as acquiring information, etc. transmitted from another device. The table, etc. (table, database, etc.) in which parameters are associated is not limited to the illustrated table, etc., and the number of parameters may be reduced or increased. Furthermore, information, etc. corresponding to parameters may be obtained using a mathematical formula, a conditional formula, etc., without using a table, etc.

[0113] The above-described embodiments are information processing devices such as the server device 10, the user terminal 20, and the operator terminal 30, and information processing systems such as the classification processing system 1 including the server device 10, the user terminal 20, and the operator terminal 30, but may also be information processing methods. The information processing methods include the same steps as those executed by the information processing system. The above-described embodiments may also be programs. The programs cause a computer to execute the same steps as those executed by the information processing system.

[0114] <Additional Notes> Furthermore, it may be provided in the following aspects.

[0115] (1) An information processing system having at least one processor, wherein the processor inputs a plurality of names to be classified into a vectorization means in an input step, the vectorization means converts the input character strings into vectors, and generates vectors whose distances are closer the closer the meanings expressed by the character strings, and the classification step classifies the plurality of names into a plurality of groups based on the distances between each of the plurality of vectors generated from the plurality of names.

[0116] According to this aspect, a plurality of names can be classified into a plurality of groups.

[0117] (2) In the information processing system described in (1) above, each of the multiple names is the name of a business operator, and in the input step, the processor inputs the multiple names as well as transaction information regarding transactions conducted by the business operators with the multiple names to the vectorization means, and in the classification step, classifies the multiple names into multiple groups based on the distance between each of the multiple vectors generated from the multiple names and the transaction information.

[0118] According to this embodiment, the accuracy of the group can be improved.

[0119] (3) In the information processing system described in (1) or (2) above, the vectorization means is an artificial intelligence module that has performed machine learning, and in the learning step, the processor causes the artificial intelligence module to learn words related to the field to which the name belongs.

[0120] According to this aspect, the accuracy of vectorization can be improved.

[0121] (4) In the information processing system described in any one of (1) to (3) above, in the classification step, the processor calculates an index value for each group that indicates the magnitude of the distance between vectors in the same group, and if the calculated index value is greater than a predetermined standard, reclassifies the group.

[0122] According to this aspect, it is possible to reduce the number of groups with low accuracy.

[0123] (5) In the information processing system described in any one of (1) to (4) above, the processor performs classification in the classification step using a method of classifying into a specified number of groups, and then reclassifies by changing the number of groups.

[0124] According to this embodiment, the number of groups can be made more appropriate.

[0125] (6) In the information processing system described in any one of (1) to (5) above, the processor, in the classification step, reclassifies a name that has been converted into a vector whose distance to other vectors in the same group is equal to or greater than a first threshold into a group among other groups whose distance between the name and the vector is closest.

[0126] According to this aspect, it is possible to reduce names that deviate significantly from the group.

[0127] (7) In the information processing system described in any one of (1) to (6) above, the processor, in a first display step, displays a name converted into a vector whose distance to other vectors in the same group is equal to or greater than a second threshold, and one or more other groups among the other groups whose distance between the name and the vector is closest, and in the classification step, reclassifies the converted name from the other groups into a group selected by the user.

[0128] According to this aspect, it is possible to reduce the effort required for the user to reclassify names.

[0129] (8) In the information processing system described in any one of (1) to (7) above, in the classification step, the processor classifies the plurality of names into a first number of groups using a method for classifying into a specified number of groups, and further classifies each group after classification into a second number of groups.

[0130] According to this embodiment, it is possible to classify the data into more appropriate groups than by classifying the data in detail all at once.

[0131] (9) In the information processing system described in (8) above, in the classification step, the processor calculates an index value for each of the first number of groups that indicates the magnitude of the distance between vectors in the same group, and sets a number corresponding to the calculated index value as the second number in each group.

[0132] According to this aspect, the degree of variation within the group can be made uniform.

[0133] (10) In the information processing system described in (8) above, in the classification step, the processor counts the number of names classified into groups for each of the first number of groups, and sets the number corresponding to the counted number of names as the second number in each group.

[0134] According to this embodiment, the size of each group can be made uniform.

[0135] (11) In the information processing system described in any one of (1) to (10) above, in a first extraction step, the processor extracts one or more words that are commonly used in the names classified into the same group, and in a second display step, displays the extracted one or more words as label candidates for the same group.

[0136] According to this aspect, it is possible to assist in determining the label.

[0137] (12) In the information processing system described in any one of (1) to (11) above, in the generation step, the processor generates a representative vector representing the vectors based on vectors generated from names that conform to a predetermined rule among the vectors generated from the multiple names, the rule being a rule that is satisfied by names that conform to existing labels, and in the classification step, classifies names that do not conform to the rule according to the distance between the vector of the name and the representative vector.

[0138] According to this aspect, it is possible to classify into groups corresponding to existing labels.

[0139] (13) In the information processing system described in any one of (1) to (12) above, the processor is configured such that each of the multiple names is the name of a business operator, each of the multiple groups is associated with a predetermined process, and in the processing step, the processor executes the process associated with the classified group for the business operator whose name is classified into that group.

[0140] According to this aspect, different processing can be performed for each group.

[0141] (14) In the information processing system described in (13) above, the processor evaluates the results of the processing in the evaluation step, and reclassifies in the classification step if the evaluation does not satisfy a predetermined result.

[0142] According to this embodiment, the processing results can be improved.

[0143] (15) An information processing method, in which a processor included in an information processing system executes each step of the information processing system described in any one of (1) to (14) above.

[0144] According to this embodiment, it is possible to promote understanding of outliers.

[0145] (16) A program that causes a computer to execute each step of the information processing system according to any one of (1) to (14) above.

[0146] According to this embodiment, it is possible to promote understanding of outliers. Of course, this is not the case. Furthermore, the above-described embodiments and modifications may be combined in any desired manner.

[0147] Finally, while various embodiments of the present invention have been described, these are presented by way of example only and are not intended to limit the scope of the invention. The novel embodiments may be embodied in various other forms, and various omissions, substitutions, and modifications may be made without departing from the spirit of the invention. The embodiments and their modifications are intended to be included within the scope and spirit of the invention, as well as within the scope of the inventions and their equivalents as defined in the appended claims. [Explanation of symbols]

[0148] 1: Classification processing system 2: Communication line 10: Server device 11: Control section 20: User terminal 21: Control unit 30: Operator terminal 31: Control unit 100: Artificial Intelligence Module

Claims

1. An information processing system including at least one processor, the processor: In the input step, a plurality of names to be classified are input to the vectorization means; The vectorization means converts an input character string into a vector, and generates vectors whose distances become closer as the meanings expressed by the character strings become closer; In the classification step, the plurality of names are classified into a plurality of groups based on distances between each of a plurality of vectors generated from the plurality of names. Information processing system.

2. 2. The information processing system according to claim 1, each of the plurality of names is the name of a business operator; the processor: In the input step, in addition to the plurality of names, transaction information regarding transactions conducted by the business operators of the plurality of names is input to vectorization means; In the classifying step, the plurality of names are classified into a plurality of groups based on distances between the plurality of names and each of a plurality of vectors generated from the plurality of names and the transaction information. Information processing system.

3. 2. The information processing system according to claim 1, the vectorization means is an artificial intelligence module that performs machine learning; the processor: In the learning step, the artificial intelligence module is made to learn words and phrases related to the field to which the name belongs. Information processing system.

4. 2. The information processing system according to claim 1, the processor: In the classification step, an index value indicating the magnitude of the distance between vectors in the same group is calculated for each group, and if the calculated index value is greater than a predetermined standard, the group is reclassified. Information processing system.

5. 2. The information processing system according to claim 1, the processor: In the classifying step, the classification is performed using a method for classifying into a specified number of groups, and the classification is performed again by changing the number of groups. Information processing system.

6. 2. The information processing system according to claim 1, the processor: In the classification step, a name converted into a vector whose distance to another vector in the same group is equal to or greater than a first threshold is reclassified into a group among the other groups whose distance between the name and the vector is closest. Information processing system.

7. 2. The information processing system according to claim 1, the processor: In the first display step, the name converted into a vector whose distance to other vectors of the same group is equal to or greater than a second threshold value and one or more other groups whose distance between the name and the vector is closest among the other groups are displayed; In the classification step, the converted name is reclassified from the other group into a group selected by a user. Information processing system.

8. 2. The information processing system according to claim 1, the processor: In the classifying step, the plurality of names are classified into a first number of groups using a method for classifying into a specified number of groups, and each group after classification is further classified into a second number of groups. Information processing system.

9. 9. The information processing system according to claim 8, the processor: In the classification step, an index value that indicates the magnitude of the distance between vectors in the same group is calculated for each of the first number of groups, and a number corresponding to the calculated index value is set as the second number for each group. Information processing system.

10. 9. The information processing system according to claim 8, the processor: In the classifying step, the number of names classified into groups is counted for each of the first number of groups, and a number corresponding to the counted number of names is set as the second number for each group. Information processing system.

11. 2. The information processing system according to claim 1, the processor: In the first extraction step, one or more words commonly used in the names classified into the same group are extracted; In a second display step, the extracted one or more words are displayed as label candidates of the same group. Information processing system.

12. 2. The information processing system according to claim 1, the processor: In the generating step, a representative vector is generated based on a vector generated from a name that conforms to a predetermined rule, among the vectors generated from the plurality of names; The rule is a rule that a name that matches an existing label satisfies, In the classification step, names that do not conform to the rules are classified according to the distance between the vector of the name and the representative vector. Information processing system.

13. 2. The information processing system according to claim 1, the processor: each of the plurality of names is the name of a business operator; A predetermined process is associated with each of the plurality of groups, In the processing step, a process associated with the classified group is executed for the business operator having the name classified into the group. Information processing system.

14. 14. The information processing system according to claim 13, the processor: In the evaluation step, a result of the processing is evaluated; In the classification step, if the evaluation does not satisfy a predetermined result, the classification is re-performed. Information processing system.

15. An information processing method, comprising: The processor of the information processing system Executing each step of the information processing system according to any one of claims 1 to 14. Information processing methods.

16. A program, A computer is caused to execute each step of the information processing system according to any one of claims 1 to 14. program.

Citation Information

Patent Citations

  • Information processor and program

    JP2014026405A