Childcare support information generation system, childcare support information generation method and program
The system addresses inefficient training support by determining non-attributable deficiencies before generating information, reducing resource consumption and ensuring accurate attribution, thus optimizing computational efficiency and guidance.
Patent Information
- Application Number
- JP2026189883
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2026-08-19
- Publication Date
- 2026-11-19
- Estimated Expiration
- 2046-08-19
AI Technical Summary
Conventional systems generate training support information based on work records, attributing deficiencies to individual members even when they are due to managerial biases or lack of opportunities, leading to inefficient resource consumption and inaccurate guidance.
An information processing system that determines whether identified deficient areas belong to a member before generating training support information, avoiding input data construction and model calls for non-attributable areas, and distinguishes between member-related and manager-related deficiencies on a terminal display.
Reduces computational resources, processing time, and input data consumption while ensuring accurate attribution of deficiencies, preventing unnecessary guidance and recording of non-member-related events.
Smart Images

Figure 0007953875000001_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present invention relates to an information processing technology for generating information that supports the development of human resources in an organization.
BACKGROUND ART
[0002] A technology is known that analyzes records created by members of an organization for business purposes, estimates the characteristics or deficiencies of the members, and generates information for development.
[0003] Patent Document 1 describes a configuration that receives support record text data input by support staff for user support, automatically generates a daily support report for the user, and further uses the same text data to analyze the types of observation items of the staff, the depth of descriptions, and observation items that are not continuously described, so as to estimate the observation characteristics and the observation deficiency area of the staff, and propose development guidance content for supplementing the observation deficiency area to the manager of the staff.
[0004] In addition, configurations are known that estimate the state of members from interview records, attendance information, self-report data and the like, predict the possibility of turnover and present intervention measures, or calculate the fitness between the capabilities of members and duties to propose personnel allocation.
PRIOR ART DOCUMENTS
PATENT DOCUMENTS
[0005]
PATENT DOCUMENT 1
SUMMARY OF THE INVENTION
PROBLEMS TO BE SOLVED BY THE INVENTION
[0006] The prior art, including the configuration described in Patent Document 1, is all based on the premise that the estimated deficiency area belongs to the member in question. That is, once a deficiency area is identified through analysis of work records, training support information targeting that member is generated and presented to the administrator.
[0007] However, the lack of descriptions appearing in work records is not necessarily attributable to the member in question. The inventor of this application confirmed the following phenomena through operation in multiple organizations.
[0008] Firstly, managerial involvement may be biased towards specific areas. If a manager fails to consistently refer to a particular area of observation and does not provide guidance, ask questions, or assign tasks related to that area, members under their management will not be provided with opportunities for experience and guidance in that area. In this case, the absence of descriptions of that area in the members' work records is not due to a lack of observation on the part of the members, but rather to a lack of opportunity.
[0009] Secondly, even with the same member, the areas appearing in the work records may change when the manager changes. When this phenomenon occurs during a period in which no change in the member's abilities is apparent, the lack of description in that area is not attributable to that member.
[0010] Conventional technology generates training support information for the member in question, even in these cases. As a result, information is presented to the manager requesting the member to correct areas where opportunities are not provided, and even if guidance is given based on this information, the lack of description is not resolved. Furthermore, because the lack is recorded in association with the member's identifier, events that do not belong to the member are accumulated as characteristics of that member.
[0011] Furthermore, generating training support information requires structuring input data for the generation model and calling the model. In the above case, the generated information cannot resolve the deficiencies in the description of the member, and therefore the computational resources, processing time, and amount of input data consumed for such generation do not lead to any results. This consumption increases cumulatively with the increase in the number of members and the number of observation areas.
[0012] The present invention aims to solve the above-mentioned problems by determining whether the identified deficient region belongs to a member prior to generating training support information, and by not configuring input data for the generation model and not calling the model for regions that are determined not to belong to a member. [Means for solving the problem]
[0013] One aspect of the present invention is an information processing system for generating training support information in an organization including multiple subjects and administrators who manage those subjects, and for displaying it on a terminal used by the administrator, comprising: a record acquisition means for storing work record data created by the subjects for their work and management record data, which is a record of responses made by the administrator to the subjects, in a storage device associated with a subject identifier and an administrator identifier; a first analysis means for classifying the work record data into multiple observation areas and identifying, for each subject, observation areas in which the amount of description in the observation area falls below a predetermined standard as deficient areas; a second analysis means for classifying the management record data into the same observation areas as the first analysis means and identifying, for each administrator, observation areas in which the number of occurrences of mentions in the observation area falls below a predetermined standard as low-involvement areas; and, prior to generating training support information, if the deficient areas of a subject are included in the low-involvement areas of the administrator associated with that subject, and the number of subjects associated with that administrator who have such deficient areas is equal to or greater than a predetermined percentage of the total number of subjects associated with that administrator, The child-rearing support information generation system comprises: an attribution determination means for classifying the deficient area as an area not attributable to a target person; a generation means for generating child-rearing support information by constructing input data for a generation model and calling the generation model for the deficient areas that were not classified as areas not attributable to a target person, and not constructing the input data for the areas not attributable to a target person and not calling the generation model; a display control means for providing a first area and a second area on the same screen of the terminal, displaying the generated child-rearing support information together with the target person's identifier in the first area, and displaying in the second area the name of the area not attributable to a target person, the number of target persons who have such an area, and a statement that no child-rearing support information has been generated for such an area, without the target person's identifier, in a display format that is identifiable from the first area; and a recording means configured to aggregate the number of times the generation means has generated child-rearing support information and the number of times child-rearing support information has not been generated due to classification as an area not attributable to a target person for each period, store them in the storage device, and output the aggregated results. [Effects of the Invention]
[0014] According to the present invention, for missing areas not belonging to a member, no input data is generated for the generation model, and the model is not called. Therefore, the number of generation process executions, processing time, and amount of input data are reduced, and the consumption of computing resources is decreased. This reduction can be output as an aggregate by the recording means. Furthermore, for areas determined not to belong, it is not associated with the subject's identifier, and it is indicated that the area appears in common for multiple subjects. Therefore, events related to the area are not recorded and presented as belonging to a specific subject. Moreover, for areas determined not to belong, no information indicating whether corrective measures are necessary for the subject is output. Therefore, guidance that cannot resolve the lack of description is not provided. [Brief explanation of the drawing]
[0015] [Figure 1] This is a block diagram showing the overall configuration of the information processing system according to this embodiment. [Figure 2] This is a flowchart showing the processing flow in this embodiment. [Figure 3] This diagram shows the classification of observation areas and the correspondence between the areas of deficiency for the subjects and the areas of low involvement for the managers. [Figure 4] This figure shows an example of a screen that the display control means displays on the terminal. [Figure 5] This flowchart shows the details of the determination process by the attribution determination means. [Figure 6] This figure shows an example of the aggregated results stored by the recording device. [Modes for carrying out the invention]
[0016] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the present embodiment, a case of application to a continuous employment support type A workplace will be described as an example, but the application target of the present invention is not limited thereto. The workplace according to the present embodiment has 1 administrator and 5 support staff members, and support records for 20 users are input on a daily basis. The support staff members input the support records from mobile terminals to support the users. The records are created for the users, and do not include information input for the self-development of the support staff members themselves. As shown in Figure 1, the information processing system 1 according to the present embodiment includes a record acquisition unit 10, a first analysis unit 20, a second analysis unit 30, an attribution determination unit 40, a generation unit 50, a display control unit 60, and a recording unit 70. The generation unit 50 is connected to an external generation model 80. The record acquisition unit 10 acquires business record data D1 and management record data D2, and stores them in a storage device 90. The business record data D1 is text data input by the support staff for supporting the users, and is stored in association with the target person identifier and the input date and time. The management record data D2 is a record of responses made by the administrator to the support staff, including comments added by the administrator to the business records, records of the administrator's statements in interviews, and business assignment history. The management record data D2 is stored in association with the administrator identifier and the target person identifier that is the recipient of the response. Each target person identifier is associated with the administrator identifier of the administrator who manages the relevant target person. When there is a change of administrator, an application start date and an application end date are attached to the association. This makes it possible to analyze the same target person by dividing the period for each administrator. The first analysis unit 20 classifies the business record data D1 into a plurality of observation areas. In the present embodiment, the observation areas are four areas: area A related to the growth and changes of the target person, area B related to physical and living conditions, area C related to operation and work efficiency, and area D related to safety and risks. However, the type and number of observation areas are not limited to these, and are set according to the nature of the business of the organization to which they are applied. The first analysis unit 20 classifies the descriptions included in the work record data D1 into the above observation regions, and calculates the occurrence count of descriptions and the description volume within a predetermined period for each target person identifier and each observation region. The first analysis unit 20 identifies an observation region where the calculated occurrence count or description volume falls below a predetermined criterion as a deficient region of the relevant target person. In the present embodiment, when the descriptions classified into the relevant observation region in the most recent 3 months account for less than 5% of the total description volume, the relevant observation region is identified as a deficient region. The relevant criterion is appropriately set according to the nature of the work of the organization. The second analysis unit 30 classifies the management record data D2 into the same observation regions as the first analysis unit 20. That is, the comments added by the manager, statements in interviews, and work assignment histories are classified into any of region A to region D. The second analysis unit 30 calculates the occurrence count of references within a predetermined period for each manager identifier and each observation region, and identifies an observation region where the relevant occurrence count falls below a predetermined criterion as a low-involvement region of the relevant manager. Here, it is important that the first analysis unit 20 and the second analysis unit 30 use the same observation regions. If the two use different classification axes, it is not possible to associate the deficiency in the target person's descriptions with the deficiency in the manager's involvement. In the present embodiment, by applying the same classification processing to both the work records and the management records, the association between the two is enabled. Prior to the generation of development support information by the generation unit 50, the attribution determination unit 40 determines, for the deficient region identified by the first analysis unit 20, whether the relevant deficient region is attributable to the target person. As shown in Figure 5, the attribution determination unit 40 classifies the relevant deficient region as a target person non-attributable region when both of the following two conditions are satisfied. The first condition is that the relevant deficient region is included in the low-involvement regions identified by the second analysis unit 30 for the manager associated with the relevant target person. The second condition is that, among the target persons associated with the relevant manager, the number of target persons having the relevant deficient region is equal to or greater than a predetermined ratio of the total number of target persons associated with the relevant manager. In this embodiment, the predetermined ratio is set to two-thirds. That is, if there are five individuals associated with the manager, the second condition is met when four or more of them have the same deficiency area. The ratio is set appropriately according to the size of the organization and the nature of the work, but if a value less than a majority is set, it is not possible to eliminate coincidences, and if a perfect match is required, the opportunities for determination are excessively limited. If both conditions are met, the deficiency is determined not to be due to insufficient observation of the subjects, but rather to a lack of opportunities related to that area due to biased involvement by the administrator. In addition to the above, the attribution determination means 40 also classifies the deficient area as a non-attributed area of the target when the administrator identifier for the same target changes and the deficient area of the target changes before and after the change. In this case, since the deficient area has changed despite no change on the target's side, the deficient area does not belong to the target. The attribution determination means 40 does not use data indicating the attributes of the subject that are not related to job suitability in its determination. Here, attributes not related to job suitability refer to matters of the subject's private life, preferences, and whether or not they have private interactions with the manager. If this data is used in the determination, there is a risk that differences not due to job-related reasons will be reflected in the attribution determination. The generation means 50 constructs input data for the generation model 80 only for the areas of deficiency identified by the first analysis means 20 that were not classified as areas not belonging to the target person by the attribution determination means 40, and calls the generation model 80 to generate childcare support information. For areas not belonging to the target person, the generation means 50 does not construct input data for the generation model 80 and does not call the generation model 80. Here, it is important to note that this configuration does not perform generation itself, rather than discarding the output after generation. In a configuration that discards the output after generation, computational resources, processing time, and the amount of input data required for structuring the input data, calling the generation model, and receiving the response are consumed. In this embodiment, these consumptions are avoided by performing the assignment determination prior to generation. As shown in Figure 4, the display control means 60 provides a first area R1 and a second area R2 on the same screen of the terminal used by the administrator. The first area R1 displays the childcare support information generated by the generation means 50 along with the identifier of the target person. The second area R2 displays, for areas not belonging to a target person, the name of the area not belonging to a target person, the number of target people in which that area appears in common, and a statement indicating that no childcare support information has been generated for that area. In the second domain R2, the subject's identifier is not displayed. This is to avoid displaying events related to this domain as belonging to a specific subject. Furthermore, information indicating whether corrective action is necessary for the subject is not displayed in the second domain R2. In other words, information indicating that guidance, warnings, or improvements should be requested from the subject is not displayed. This is because, since this domain has been determined not to belong to the subject, information requesting corrective action from the subject cannot compensate for the lack of description. In the second area R2, the frequency of mentions in the management record data D2 corresponding to the area not attributable to the subject may be displayed in a format that allows comparison with other observation areas. The display control means 60 causes the first area R1 and the second area R2 to be displayed in a distinguishable display format. Specifically, the background color, border lines, or headings are made different. This allows the administrator to distinguish and recognize matters attributable to the subject from matters not attributable to the subject. The recording means 70 aggregates the number of times the generation means 50 generated childcare support information, the number of times generation was not performed due to classification into an area not belonging to the target person, and the number of calls to the generation model 80 and the amount of input data reduced as a result, for each period, and stores them in the storage device 90. The recording means 70 is configured to output the stored aggregated results. Figure 6 is an example of the display of the aggregated results. This aggregation allows for a quantitative understanding of the amount of computing resources saved by introducing this system. [Examples]
[0017] In the operation of this embodiment, of the areas of deficiency identified for the five support staff members, on average about 30% were classified as areas not attributable to the target individual. In this case, both the number of calls to the generative model 80 and the amount of input tokens are reduced by a similar amount. Let's look at a specific example. In a business office where five support staff members, P1 to P5, are employed under manager M, the first analysis means 20 identifies area C as a deficient area for four of the staff members, P1, P2, P3, and P4. The second analysis means 30 analyzes manager M's management record data and finds that the number of mentions of area C in the most recent three months is 3, which is below a predetermined standard compared to area A (42 mentions), area B (38 mentions), and area D (21 mentions). Therefore, area C is identified as a low-involvement area for manager M. The attribution determination means 40 classifies the deficient areas related to area C from P1 to P4 as areas not attributable to the subject, because they are included in the low-involvement area of administrator M, and because 4 out of the 5 subjects associated with administrator M have the deficient areas, representing more than two-thirds of the total. As a result, the generation means 50 does not construct input data for the generation model 80 for areas C from P1 to P4, and does not perform a call. On the other hand, the missing area related to area D identified for P5 is not included in the low-involvement area of administrator M, and therefore is not classified as an area not belonging to the target person, and training support information is generated. The display control means 60 displays the nurturing support information related to P5 in the first domain R1 along with the subject's identifier, and in the second domain R2, it displays the name of domain C, the number of subjects common to that domain (4 people), and a statement that no nurturing support information has been generated for that domain, without the subject's identifier. As another example, consider the case where, for support staff member P6, area C was identified as a deficient area during the period under manager M, but after transferring to manager N, area C was no longer identified as a deficient area. In this case, the attribution determination means 40 detects that the deficient area has changed before and after the change in manager identifier, and classifies the deficient area as an area not attributable to the target person. [Industrial applicability]
[0018] This invention is applicable to any organization, regardless of industry, where work records are continuously created and managers respond to those who create such records. For example, when applied to a tax accounting firm, observation areas can be set by business field. In this case, if the involvement of the manager (the firm's director) is biased towards a particular field, descriptions of areas other than that field will not appear in the records of the employees under their management. According to the configuration of this invention, such deficiencies are determined not to be attributable to the employees, and training support information targeting employees is not generated. In addition, it can be applied to medical institutions, educational institutions, construction companies, and other organizations that perform work involving record-keeping. [Explanation of Symbols]
[0019] 1. Information Processing System 10. Record acquisition means 20 First analysis method 30 Second analysis means 40. Attribution determination means 50 Generation means 60 Display control means 70 Recording means 80 Generative Models 90 Storage device D1 Business Record Data D2 Management Record Data R1 First area R2 second area
Claims
1. An information processing system that generates training support information for an organization including multiple individuals and the administrator who manages those individuals, and displays it on a terminal used by the administrator, A record acquisition means that stores work record data created by the subject for work purposes and management record data, which is a record of responses made by the administrator to the subject, in a storage device, associating them with the subject identifier and the administrator identifier. A first analysis means classifies the aforementioned work record data into multiple observation areas and identifies, for each subject, observation areas in which the amount of description falls below a predetermined standard as deficient areas. A second analysis means classifies the management record data into the same observation areas as the first analysis means, and for each administrator, identifies observation areas where the number of mentions in the observation area falls below a predetermined standard as low-involvement areas. Prior to generating training support information, an attribution determination means classifies the aforementioned deficient area of a target person as a non-attributed area of the target person if the deficient area of the target person is included in the aforementioned low-involvement area of the manager associated with the target person, and the number of target persons associated with the manager who have the deficient area is equal to or greater than a predetermined percentage of the total number of target persons associated with the manager. Regarding the aforementioned insufficient areas that were not classified as areas not belonging to the target person, the generation means generates support information for child-rearing by constructing input data for the generation model and calling the generation model, while for the areas not belonging to the target person, the generation means does not construct the input data and does not call the generation model. The terminal has a display control means that provides a first area and a second area on the same screen, the first area displays the generated childcare support information along with the subject's identifier, and the second area displays the name of the area not belonging to the subject, the number of subjects who have said area, and a statement that no childcare support information has been generated for said area, without the subject's identifier, in a display format that is distinguishable from the first area. A recording means configured to aggregate the number of times the generation means generated nurturing support information and the number of times it did not generate nurturing support information because it classified the target person into the non-assigned area, for each period, and store this aggregated information in the storage device, and to output the aggregated results, A system for generating information to support child development, characterized by comprising the following features.
2. The child development support information generation system according to claim 1, characterized in that the display control means does not display information indicating whether or not corrective measures are necessary for the subject in the subject non-associated area in the second area.
3. The training support information generation system according to claim 1, characterized in that when the administrator identifier associated with the same target person is changed, and the deficient area of the target person changes before and after the change, the deficient area is classified as a deficient area not associated with the target person.
4. The child development support information generation system according to claim 1, characterized in that the observation area includes an area relating to the growth and changes of the subject, an area relating to the physical and living conditions, an area relating to operation and work efficiency, and an area relating to safety and risk.
5. The child-rearing support information generation system according to claim 1, characterized in that the display control means displays in the second region the number of occurrences of mentions in the management record data corresponding to the non-assigned subject region in a format that can be compared with the number of occurrences in other observation regions.
6. The child-rearing support information generation system according to claim 1, characterized in that the recording means calculates and outputs the amount of reduction in the number of calls to the generation model and the amount of input data due to not generating child-rearing support information by classifying the subject into the area not belonging to the subject.
7. The nurturing support information generation system according to claim 1, characterized in that the attribution determination means does not use data indicating the subject's private life, preferences, and whether or not they have private interactions with the administrator in classifying the subject's non-attribution area.
8. The training support information generation system according to claim 1, characterized in that the record acquisition means acquires, as management record data, comments added by the manager to the work record, records of statements made by the manager during interviews, and the history of work assignments.
9. A method for generating training support information for an organization that includes multiple individuals and administrators who manage those individuals, and displaying it on a terminal used by the administrator, wherein a computer... A record acquisition step involves storing work record data created by the subject for work purposes and management record data, which is a record of responses made by the manager to the subject, in a storage device, associated with the subject identifier and the manager identifier. The first analysis step involves classifying the aforementioned work record data into multiple observation areas and identifying, for each subject, observation areas where the amount of description falls below a predetermined standard as deficient areas. A second analysis step involves classifying the management record data into the same observation areas as in the first analysis step, and for each administrator, identifying observation areas where the number of mentions in that observation area falls below a predetermined standard as low-involvement areas. Prior to generating training support information, if the aforementioned deficient area of the target person is included in the aforementioned low-involvement area of the manager associated with the target person, and the number of target persons associated with the manager who have the deficient area is equal to or greater than a predetermined percentage of the total number of target persons associated with the manager, the deficient area is classified as a non-associated area of the target person. For the areas of the aforementioned deficiency that were not classified as areas not belonging to the target person, the generation step involves generating input data for the generation model and calling the generation model to generate support information, while for the areas not belonging to the target person, the generation step does not involve generating input data and not calling the generation model. A display control step provides a first area and a second area on the same screen of the terminal, the first area displays the generated childcare support information along with the subject's identifier, and the second area displays the name of the area not belonging to the subject, the number of subjects who have said area, and a statement that no childcare support information has been generated for said area, without the subject's identifier, in a display format that is distinguishable from the first area. A recording step which aggregates the number of times the development support information was generated in the generation step and the number of times the development support information was not generated because the target person was classified into the non-assigned area, for each period, and stores them in the storage device, A method for generating training support information, characterized by performing the following actions.
10. A program for causing a computer to execute the method for generating childcare support information described in claim 9.
Citation Information
Patent Citations
Labor situation management system
JP2022138227A
Business support document creation device
JP2025177897A
System
JP2026101261A
Information processing device, information processing system, and information processing method
JP7890001B1
Information processing apparatus, information processing method, and non-transitory computer-readable medium
US20260011260A1
