JP6980608B2 - Examination support device, examination support method, and service provision method - Google Patents

Description

The present invention relates to an examination support device, an examination support method, and a service providing method.

Conventionally, many other examination operations such as building confirmation, immigration control, customs examination, business license, tax examination, loan examination, etc. are known. In such examination work, the examiner examines whether or not the content of the application satisfies the predetermined requirements.

Patent Document 1 discloses a system that supports such examination work. This conventional system estimates whether the applicant's personal attribute information (nationality, gender, age) corresponds to the past tendency of the violator by machine learning technology, and the estimation result is used for examination work. offer.

Japanese Unexamined Patent Publication No. 2017-4328

An object of the present invention is to provide an examination support technique that appropriately supports examination work by adopting an element different from Patent Document 1 (personal attribute information) and estimating a requirement deficiency.

In order to solve the above-mentioned problems, one of the representative inventions is an acquisition unit that acquires information on an application matter, and applies the acquired information on the application matter to a judgment criterion to obtain an estimation result of deficiency of requirements. It is equipped with an estimation department and a support department that provides information on the estimation results of deficiencies in the application to the examination work of the application and supports the examination work.

The criteria used here are
Input: Attribute information of applicant, information of examination process Learning target: It is the result of machine learning that at least the history of deficiency of application requirements is used as training data.

By adopting the information on the examination process as a new element, it becomes possible to estimate deficiencies in requirements from a new perspective and appropriately support the examination work.

Issues, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

Configuration diagram of the examination business system. A sequence diagram illustrating the overall processing of the examination business system. The sequence diagram explaining the accumulation operation of the examination database 30. Explanatory drawing of applicant attribute information 31. Explanatory drawing of examination work history 32. Explanatory drawing of rule violation case information 33. The sequence diagram explaining the learning process of the determination criterion 48. This is an example of the calculation method of the newly added attribute 441. This is an example of the newly added attribute 441. This is an example of the determination criterion 48. It is a display screen of a judgment standard. A sequence diagram (1/2) explaining the examination support. A sequence diagram (2/2) explaining the examination support. It is a display screen of the detailed examination distribution result. This is an example of the relationship between the revision of the legal system and the date of examination and disposition. This is an example of the definition of dissimilarity between the legal system and the date of disposal. This is a display example of a violation case. This is an example in which the information on the examination process is converted into an intermediate node.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the drawings.

1. 1. Explanation of Configuration FIG. 1 is a diagram illustrating the configuration of an examination business system (including an examination support device).

The examination terminal 10, the examination server 20, the examination database 30, and the support server 40 are provided in this system so as to be able to communicate with each other via the network 5.

The examination terminal 10 is a client terminal for the examiner 1 (operator) to operate the examination work, and since a plurality of examiners 1 perform the examination work at the same time, a necessary number of units are prepared. The examination terminal 10 includes an interface 11, a CPU 12, a main memory 13, an input device 15, and a display device 16. The main memory 13 stores software for executing the examination work client function 14 on the CPU 12.

The examination server 20 provides services such as data processing and data management for examination work to the examination terminal 10. The examination server 20 includes an interface 21, a CPU 22, and a main memory 23. The main memory 23 stores software for executing the examination business server function 24 by the CPU 22.

The examination database 30 sequentially accumulates the history of the work processed by the examination terminal 10 and the examination server 20. For example, the examination database 30 accumulates and retains the applicant attribute information 31, the examination work history 32, and the rule violation case information 33.

The support server 40 provides a support service for examination work to the examination server 20 and the examination terminal 10. The support server 40 includes an interface 41, a CPU 42, a main memory 43, and a storage device 47. The main memory 43 stores software for the CPU 42 to perform the function 44 for creating the determination criterion 48 as to whether or not detailed examination is required. The determination criterion 48 created here is stored in the storage device 47. Further, in the main memory 43, software for executing the function 45 for allocating whether or not the application matter requires detailed examination is stored in the CPU 42. Further, the main memory 43 stores software for the CPU 42 to execute the function 46 for extracting past violation cases for the application cases assigned to the detailed examination.

2. 2. Flow of overall processing FIG. 2 is a sequence diagram illustrating the overall processing of the examination business system.
Hereinafter, the flow of the entire process will be described along with the step numbers in FIG.

Step S1: The examiner 1 turns on the power of the examination terminal 10 and activates the software of the examination business client function 14. By the login process of the examination work client function 14, the examination terminal 10 cooperates with the examination server group (examination server 20, examination database 30, and support server 40), and the subsequent examination work is started.

Step S2: The examination database 30 acquires the information exchanged in the system according to the progress of the examination work as a history, classifies and stores the applicant attribute information 31, the examination work history 32, and the rule violation case information 33. ..

Step S3: The examination terminal 10 requests the support server 40 in a timely manner for the determination criterion 48 for allocating the application matter to the detailed examination matter / normal examination matter.

Step S4: The support server 40 transmits the machine-learned determination criterion 48 to the examination terminal 10 in response to the request of the examination terminal 10. The details of machine learning will be described later.

Step S5: The support server 40 supports the examination work of the examiner 1 in cooperation with the examination server 20, the examination database 30, and the examination terminal 10. The contents of support here include a process of sorting application cases according to the determination criterion 48, a process of extracting inferred violation cases and displaying them on the examination screen.

Step S6: When the examination disposition of the application matter by the examiner 1 is completed, the examination terminal 10 transmits the processing data related to the examination disposition to the system for receiving the approval of the approver and the system for notifying the applicant of the examination disposition. , Complete the examination work. After that, the examination terminal 10 cancels the cooperation with the examination server group (examination server 20, examination database 30, and support server 40) by the logout process of the examination business client function 14.

3. 3. Details of History Accumulation FIG. 3 is a sequence diagram illustrating the history accumulation of the examination database 30.
Hereinafter, the details of the history accumulation (FIG. 2S2) described in the overall processing will be described along with the step numbers in FIG.

Step S100: The examination terminal 10 logs in to the examination work system and requests the examination server 20 to accumulate data of the application matter and the examination work.

Step S101: The examination server 20 applies for information input from a plurality of examination terminals 10 regarding an application matter and examination work, and information acquired by the examination server 20 from an external cooperation system (not shown) in the examination database 30. Person attribute information 31, examination work history 32, and rule violation case information 33 are classified and accumulated.

Hereinafter, the accumulated data in the examination database 30 will be described with reference to FIGS. 4 to 6.
FIG. 4 illustrates the applicant attribute information 31 in the examination database 30. The applicant attribute information 31 stores the applicant ID 310 and the attribute information (attribute 311 such as gender / date of birth / address / occupation and its value 312) as information for each individual applicant. These information are information input by the applicant on paper or online application form, and are stored in the examination database 30 through the input operation of the examination terminal 10 and the information collection of the examination server 20. At this time, it is preferable that the examination database 30 performs name identification processing on the applicant attribute information 31 so that the attribute information of the same applicant does not overlap.

FIG. 5 exemplifies the examination work history 32 in the examination database 30. The examination work history 32 holds the process of examination (examination type, work history, date by time stamp, etc.) as information for each applicant ID. This information is automatically generated according to the examination work performed by the examiner 1 on the examination terminal 10, and is stored in the examination database 30 via the processing of the examination server 20.

In the example shown in FIG. 5, the applicant with the applicant ID “1” applied for the certification examination of qualification A, and the examiner accepted the application (history 3201) and the process of disposing of the certification examination (history). 3202) is recorded. In addition, the process in which the applicant applied for the renewal examination of qualification A and the examiner accepted the application (history 3203) and the process of disposing of the renewal examination (history 3204) are recorded. Furthermore, the process in which the applicant applied for the change examination from qualification A to qualification B and the examiner accepted the application (history 3205) and the process of disposing of the change examination (history 3206) are recorded. ..

On the other hand, after the applicant with the applicant ID "2" goes through the examination process (history 3207 to 3210), the history 3211 in which the applicant's rule violation is found and the qualification is disqualified is recorded. Whether or not the applicant has a history of rule violation can be determined by the presence or absence of the history of disqualification 3211.

FIG. 6 illustrates the rule violation case information 33 in the examination database 30. The rule violation case information 33 holds the violation case 331 (procedure type, violation content, specific content of the violation) as information for each applicant ID of the applicant who violated the rule. This information is stored in the examination database 30 through a process in which the examiner 1 inputs the information from the examination terminal 10 or the examination server 20 acquires and registers the violation disposition from the outside.

Step S102: The examination terminal 10 acquires the history (registration result) accumulated in the examination database 30 in a timely manner and displays it on the examination screen.

By performing the above operations (FIGS. 3S100 to S102) between the plurality of examination terminals 10 and the examination server 20, the amount of training data required for machine learning is sequentially accumulated in the examination database 30. To.

4. Details of Learning Process FIG. 7 is a sequence diagram illustrating the learning process of the determination criterion 48.
Hereinafter, the details of the machine learning (FIGS. 2S3 to S4) mentioned in the overall processing will be described along with the step numbers in FIG.

Step S200: The examination terminal 10 requests the support server 40 in a timely manner for the determination criterion 48 for allocating the application matter to the detailed examination matter / normal examination matter.

Step S201: The support server 40 accesses the examination database 30 and acquires the examination work history 32.

Step S202: The support server 40 extracts the time-series information of the examination work history 32 into the applicant ID unit and the examination process unit, and generates the examination process information based on the extracted examination process.

As the information on the examination process here, information having either or both of the following properties is preferable.
(1) Information on the examination process that shows common tendencies (for example, certain similar or specific tendencies) among past violators (2) Information on the examination process that tends to change depending on the situation of deficiencies in application requirements

Further, it is preferable to adopt any one type or a combination of two or more types of the following information as the information of the examination process having this kind of property.
(A) Length between time stamps of examination process (B) Number of examination process (C) Type of examination process (including time series pattern of type)

Further, as the information on the examination process, any one type or a combination of two or more types of the following information is more preferable.
(A) Background information about the period required from the acceptance of the application or the start of the examination to the examination disposition (b) Background information about the period from the first examination disposition to the final (recent) examination disposition (c) About the number of times the application is updated Background information

FIG. 8 is a diagram showing a calculation process of the newly added attribute 441 as an example of information on the examination process.
FIG. 9 is a diagram showing the calculation result of the newly added attribute 441.
When the support server 40 detects the process of acceptance of the qualification A certification application for the applicant ID "1" from the examination work history 32, the shortest qualification A examination disposition history 3202 after the time stamp of this acceptance history 3201. The time stamp of is detected, the period 4410 (the period between the examination history) between the two time stamps is detected, and it is recorded in the newly added attribute 441 together with the label of the required period of the certification examination of the qualification A.

Further, when the support server 40 detects the process of acceptance of the application for renewal of qualification A for the applicant ID "1" from the examination work history 32, the examination for the shortest renewal of qualification A after the time stamp of the acceptance history 3203 is performed. The time stamp of the disposal history 3204 is detected, the period 4411 (the period between the examination history) between the two time stamps is detected, and the time stamp is recorded in the newly added attribute 441 together with the label of the required period of the renewal examination of the qualification A.

Further, when the support server 40 detects the process of acceptance of the qualification change application (from A to B) for the applicant ID "1" from the examination work history 32, the shortest change after the time stamp of the acceptance history 3205 is detected. Detects the time stamp of the examination disposition history 3206, detects the period 4412 (the period between the examination history) between the two time stamps, and adds a new one with the label of the change examination period from qualification A to qualification B. Record in attribute 441.

Further, when the support server 40 detects the background of the first examination disposition for the qualification A for the applicant ID "1" from the examination work history 32, the support server 40 determines the qualification A after the time stamp of the first examination disposition history 3202. The time stamp of the final examination disposition history 3206 is detected, the period 4413 (the period between the examination history) between the two time stamps is detected, and the time stamp is recorded in the newly added attribute 441 together with the label of the certification period of qualification A.

Further, the support server 40 counts the number of updates of the qualification (application) 4414 for the applicant ID "1" from the examination work history 32, and records it in the newly added attribute 441 together with the label of the number of updates of the application.

The support server 40 creates the information 441-4418 shown in FIG. 9 by performing such processing for each applicant ID, and holds it as a history in the newly added attribute 441. At this time, by rearranging the information 441-4418 in chronological order, the information of the type of examination process (including the time-series pattern of the type) is also retained.

Step S203: Subsequently, the support server 40 acquires the history of the personal attribute information of the applicant from the applicant attribute information 31 shown in FIG.

Step S204: The support server 40 associates the applicant attribute information 31 (applicant attribute information) and the newly added attribute 441 (examination process information) with a common applicant ID, and then integrates them as training data. Data and array for batch processing. These data are inputs (state variables) such as branching of judgment criteria for machine learning and nodes.

The support server 40 assigns teacher data and reward data to individual columns of training data so that deficiencies in application requirements are targeted for learning. The learning target here is defined as a learning goal such as a teacher role (on the other hand, a teacher role), a reward (deduction), a correct answer (incorrect answer), a model (non-model), and a gain (loss) in machine learning. Is. In addition, this learning target becomes an estimation answer by selecting the possibility of the estimation result or the estimation result that maximizes the possibility in the judgment criteria after machine learning.

Insufficient requirements can be defined in various ways, such as "became a violator such as disqualification", "re-examination", and "application was rejected". It is preferable to appropriately design which definition is included in the requirement deficiency from the viewpoint of the system, such as to what extent it is useful to provide the possibility of the requirement deficiency as information for examination support.
Further, the support server 40 may classify deficiencies of requirements into violation provisions or representative cases, and assign teacher data and reward score data to individual columns of training data so that the classification code becomes a learning target. ..

Step S205: The support server 40 performs machine learning using the created training data, and creates a determination criterion 48.

Machine learning here includes decision tree learning, correlation learning, neural network learning, machine learning using the training target of training data as a teacher value, learning by deep learning error back propagation method, and other machines. General learning techniques can be applied.
For example, in decision tree learning, branch rules are created hierarchically so that the information gain (decrease in purity of the learning target due to branching) gradually increases.

FIG. 10 is an example of the criterion 48 obtained as a result of this decision tree learning. The determination criterion 48 includes distribution criteria 481 to 483 described by explanatory variables that are characteristic of the decision tree. By satisfying any of the distribution criteria 481 to 483, it is estimated that there is a high possibility that the requirements are deficient. Further, when the distribution criteria 481 to 483 overlap, it may be possible to obtain an estimation result that the larger the number of distribution criteria that are satisfied at the same time, the higher the possibility that the requirements are deficient.
The support server 40 stores the determination criterion 48 created in this way in the storage device 47.

Step S206: The examination terminal 10 displays the content of the judgment criterion 48 acquired from the support server 40 on the screen of the examiner as shown in FIG.

As the determination criterion 48 of the decision tree, the distribution criteria 481 to 483 are displayed by explanatory variables. Therefore, there is an advantage that the contents of the distribution criteria 481 to 483 are easy for the examiner to understand, and the examiner can easily understand the result of the distribution according to the distribution criteria 481 to 483.

5. Details of Examination Support FIGS. 12 and 13 are sequence diagrams illustrating examination support.

Hereinafter, the details of the examination support (FIG. 2S5) described in the overall processing will be described along with the step numbers of FIGS. 12 and 13.

Step S300: The examination terminal 10 inquires the examination server 20 of the application matter for which information has been input by the applicant or the examiner.

Step S301: The examination server 20 requests the support server 40 to determine whether or not the inquired application item meets the determination criterion 48.

Step S302: The support server 40 extracts the personal attribute information of the examiner from the information of the application matter. The support server 40 searches the past applicant attribute information 31 for the applicant's attribute information, and extracts the ID of the applicant corresponding to the past (hereinafter referred to as the corresponding person ID). If the corresponding person ID is not found, the support server 40 may extract the ID of the past applicant having a certain degree of similarity to the applicant as the corresponding person ID.

Step S303: The support server 40 uses the relevant person ID as a search key to search for the applicant ID of the examination work history 32, and extracts the background of the past examination in the relevant person ID. The extracted past examination process is processed in the same manner as the training data creation procedure described above, and information on the examination process is obtained.

Step S304: The support server 40 collates the personal attribute information of the relevant person ID with the determination standard 48 by inputting (estimated state variable) the personal attribute information of the corresponding person ID and the information of the past examination process of the corresponding person ID, and "determination criteria 48". Matching result (estimation result of deficiency of requirements) ”and“ corresponding number of distribution standard ”are returned to the examination terminal 10 via the examination server 20.

Step S305: When the application matter meets the determination criterion 48, the examination server 20 shifts the operation to the extraction of the violation case (step S307).

Step S306: The examination terminal 10 determines that there is a high possibility that the requirements are inadequate, and allocates the application case that meets the judgment criterion 48 to the detailed examination case. On the other hand, the examination terminal 10 determines that there is a low possibility that the requirements are deficient for the application case that does not meet the judgment criterion 48, and sorts the application case into a normal examination case. As shown in FIG. 14, the sorting results of this detailed examination case are "matching result of judgment standard 48 (estimation result of deficiency of requirements)", "corresponding number of sorting standard", and "explanation of contents of matching sorting standard". Is displayed on the screen to inform the examiner.

It should be noted that this distribution process can also be collectively processed for a plurality of examination cases scheduled by the examiner. In that case, in addition to displaying the detailed examination case / normal examination case for each examination case, it is possible to display the number of cases such as "detailed examination case ○ case / normal examination case ○ case". According to this number display, the examiner can systematically allocate the time required for the examination of the application matter.

Step S307: The examination server 20 requests the support server 40 for a past violation case having an attribute similar to that of the applicant (application matter) this time.

Step S308: The support server 40 extracts past cases stored in the rule violation case information 33 whose matching status of the determination criterion 48 is similar to that of the current application case. When extracted in this way, past projects (hereinafter referred to as similar projects) with a high degree of similarity to the attributes of the application project (applicant attribute information, examination history information) are extracted using the judgment criterion 48 as a similar scale. ..
For example, in the case of decision tree analysis, past projects belonging to the group to which the application project is conditionally branched are similar projects.
Violation cases of this similar case will be useful reference information when the examiner considers the deficiency of the requirements of the application case.

Step S309: As shown in FIG. 15, the judgment of past violation cases follows the provisions of the legal system 3220 to 3222 and the legal rules at the time of the judgment. This provision will be revised in a timely and progressive manner according to social conditions. Therefore, as the judgment period 3202 to 3210 is revised over and over again, past violation cases will no longer be useful for examination. Therefore, as shown in FIG. 16, the support server 40 accumulates the number of generations of the revision included in the time difference of the judgment time to obtain the dissimilarity. The support server 40 reflects the revised status of the provisions in the degree of similarity by lowering the degree of similarity of applicants for similar projects according to the degree of dissimilarity.

Step S310: The support server 40 extracts a predetermined number of violation cases from the violation cases of similar cases in descending order of the degree of similarity.

Step S311: The violation case extracted by the support server 40 is transmitted to the examination terminal 10 via the examination server 20.

Step S312: The examination terminal 10 displays the received violation case on the screen as shown in FIG. 17 and provides information to the examiner.

6. Supplementary matters of the example In the example, the past applicable person corresponding to the applicant is searched, and the information of the past examination process of the applicable person is used as a learning element. It is also preferable to replace this part with a machine learning rule. That is, as shown in FIG. 18, the support server 40 applies the attribute information of the applicant to the first rule 701 and obtains the estimation result of the information on the examination process. Further, the support server 40 applies the estimation result of the applicant's attribute information and the examination process information to the second rule 702, and obtains the estimation result of the requirement deficiency or the violation case. This first rule 701 is the result of machine learning in which the attribute information of the applicant is input and the history data for which the information of the examination process is the learning target is at least the training data. Further, the second rule 702 is the result of machine learning in which the attribute information of the applicant and the information of the examination process are input, and the historical data for learning the deficiency of requirements and the violation cases is at least the training data. By adopting the examination process information to the intermediate node in this way, it is possible to estimate even a new applicant who does not have the process information using the examination process information as a learning element.

In the embodiment, the examination server 20, the examination database 30, and the support server 40 are each distributed and configured by a plurality of server computers, but it is also preferable to appropriately configure them integrally.

Further, in the embodiment, the period from the acceptance of the application to the disposition of the examination is calculated as the information on the examination process. However, if there is a predetermined delay from acceptance to the start of examination due to waiting for examination, this delay is not related to the situation of deficiency of requirements on the applicant side. Therefore, the period from the start of the examination to the disposition of the examination may be calculated as information on the examination process.

In the embodiment, as the information on the examination process, a time-series pattern of the examination process (such as changing from the certification disposition of qualification A to qualification B in an unnatural short period of time) may be adopted. By using the time-series pattern of such a process as an estimation element, machine learning about "the tendency of the time-series pattern of the examination process", which is a requirement deficiency, becomes possible, and more appropriate examination support becomes possible.

In addition, it is preferable that the applicant includes a person to whom the applicant delegates a proxy (hereinafter referred to as an application proxy). Even for the application agent, there may be a certain tendency between the background of the examination and the deficiency of the requirements, so there is a certain effect on the examination support.

Furthermore, the application agent represents multiple applicants, and the past examination process is likely to be accumulated. Therefore, it is suitable for examination support that employs information on the examination process.

In addition, when there are multiple application agents or a corporation, machine learning and estimation processing that raises the priority of the actual application agent such as the first person or the person in charge is also suitable.

In the embodiment, the dissimilarity is defined by accumulating the number of generations of revision included in the period difference from the judgment time of the past case to the present time. The number of generations of this amendment may be defined as the number of generations as large as the major amendment that affects the examination, and the number of generations of the amendment that does not affect the examination may be substantially defined as zero. This substantive definition of the number of generations makes it possible to define the degree of dissimilarity based on the substantive impact of the amendment.

Further, in the embodiment, the estimation result of the deficiency of the requirement is provided as a service to the examination work, but the present invention is not limited to this. For example, the support server 40 may provide a service of providing the estimation result of the deficiency of requirements to the applicant via a communication means addressed to the applicant. With such a service, the applicant can voluntarily correct or cancel the application matter upon receiving a warning of deficiency of requirements. Such services not only support applicants, but also promote voluntary correction and cancellation of deficiencies in requirements, resulting in support for application acceptance and examination operations.

7. Effects of Examples The effects of general trends in Examples will be described below. In machine learning, judgment criteria are appropriately created so as to match the tendency of the learning element and the learning target. Therefore, in a special type of examination, the result of machine learning may differ from the general tendency described here. For such special types of screening, the results of machine learning take precedence.

(1) The circumstances that occur in the examination are events that reflect the deficiencies in the requirements of the application. Therefore, by adopting the judgment criteria machine-learned using the information of the examination process as a new learning element, it becomes possible to estimate the deficiencies in the requirements from a new viewpoint and appropriately implement the examination support.

(2) In the embodiment, the attribute information of the applicant and the information on the examination process are combined to form a new learning element. Therefore, the estimation of deficiencies in requirements will be further detailed, and it will be possible to appropriately implement examination support.

(3) In the embodiment, as the information on the examination process, the information on the examination process showing a certain similar tendency in the past violators is adopted as the learning element. Therefore, it is possible to determine whether or not the person has the same tendency as the past violator (whether or not there is a high possibility of becoming a violator) in a format inferred from the information on the examination process.

(4) In the embodiment, as the information on the examination process, the information on the examination process, which tends to change depending on the situation of the deficiency of the application requirements, is adopted as the learning element. Therefore, it is possible to determine the possibility of inadequate requirements in the form of inferring the cause inherent in the application by tracing backward from the information on the examination process.

(5) In the embodiment, as the information on the examination process, the process information on the period from the acceptance of the application (or the start of the examination) to the disposition of the examination is adopted as a learning element. Generally, the longer the period required for the examination, the more likely it is that the examiner will have a difficult examination target and have deficiencies in the requirements. Therefore, it is possible to determine that the longer the required period of this examination, the higher the possibility that the requirements will be deficient.

(6) In the embodiment, as the information on the examination process, the process information on the period from the first examination disposition to the final (recent) examination disposition is adopted as a learning element. In general, such a long qualification period is presumed to have been in a stable state without being dismissed from the qualification on the way. Therefore, the longer this qualification period is, the less likely it is that the requirements will be inadequate.

(7) In the embodiment, as the information on the examination process, the process information on the number of updates of the application is adopted as the learning element. In general, it can be inferred that the fact that the number of renewals is large means that the company was in a stable state without being disqualified. Therefore, it is possible to determine that the greater the number of updates, the lower the possibility that the requirements will be deficient.

(8) In the embodiment, the application cases are sorted into detailed examination cases based on the estimation result of the deficiency of requirements. According to this distribution process, the examiner can examine in detail the examination cases that are likely to have deficiencies in the requirements, so that the quality of the examination can be improved. In addition, the speed of examination can be efficiently increased by conducting regular examinations for examination cases that are unlikely to have deficiencies in requirements. As a result, it becomes possible to solve the conflicting problems of quality and speed of examination in a well-balanced manner.

(9) In the embodiment, a violation case of deficiency of requirements is requested for a past case (similar case) similar to the application case, and the violation case of the similar case is provided to the examination work. Therefore, the examiner can improve both the quality and speed of the examination by giving priority to the examination and examination of the application case for this violation case.

(10) In the embodiment, the attribute information of the applicant and the information on the examination process are combined to extract violation cases of similar cases. Therefore, it is possible to extract more appropriate violation cases while further narrowing down the number of violation cases to be extracted.

(11) In the example, the degree of similarity of the past projects before the revision is lowered. As a result, it is possible to extract appropriate violation cases in consideration of the revision status of the regulations.

(12) In the embodiment, the greater the number of revisions between the past project and the application project, the lower the degree of similarity. As a result, more appropriate violation cases can be extracted based on the progressive revision status of the regulations.

5 ... Network, 10 ... Examination terminal, 11 ... Interface, 12 ... CPU, 13 ... Main memory, 14 ... Examination business client function, 15 ... Input device, 16 ... Display device, 20 ... Examination server, 21 ... Interface, 22 ... CPU, 23 ... main memory, 24 ... examination work server function, 30 ... examination database, 31 ... applicant attribute information, 32 ... examination work history, 33 ... rule violation case information, 40 ... support server, 41 ... interface, 42 ... CPU, 43 ... main memory, 47 ... storage device, 48 ... judgment criteria, 701 ... first rule, 702 ... second rule

Claims (10)

The acquisition department that acquires the information of the application matter, and
An estimation unit that applies the acquired information on the application to the judgment criteria to obtain the estimation result of deficiency of requirements, and
It is equipped with a support department that provides information on the estimation results of the deficiencies in the requirements to the examination work of the application case and supports the examination work.
The criterion used by the estimation unit is
An examination support device characterized in that it is the result of machine learning in which the attribute information of the applicant and the information on the examination process are input, and the history of the deficiencies in the requirements as the learning target is at least the training data. In the examination support device according to claim 1,
Information on the examination process is available.
Information on the examination process showing common tendencies among past violators,
Information on the examination process, which tends to change depending on the situation of deficiencies in the application requirements,
Examination support device characterized by containing at least one of. In the examination support device according to claim 2,
Information on the examination process is available.
Information on the process from the acceptance of the application or the start of the examination to the disposition of the examination,
Background information about the period from the first examination disposition to the final examination disposition,
Background information about the number of renewals of the application and
Examination support device characterized by containing at least one of. In the examination support device according to any one of claims 1 to 3,
The support department
An examination support device characterized by allocating the application items to detailed examination items based on the estimation result of the deficiency of requirements. In the examination support device according to any one of claims 1 to 4,
The estimation unit
The degree of similarity between the information of the application matter and the information of the past matter for which the application matter had deficiency was obtained, and the case of the deficiency of the requirement for the past matter having a high degree of similarity was obtained.
The support department
An examination support device characterized by providing information on cases of deficiencies in the requirements to the examination work of the application case and supporting the examination work. In the examination support device according to claim 5,
The estimation unit
An examination support device characterized in that the degree of similarity is lowered for past projects with different applicable provisions based on the information on the revision time of the provisions applied to the examination of the application matter. In the examination support device according to any one of claims 1 to 6.
The criterion used by the estimation unit is
The first rule, which is the result of machine learning that inputs the attribute information of the applicant and uses the history of the examination process as the learning target at least as training data,
It has a second rule that is the result of machine learning that inputs the attribute information of the applicant and the information of the examination process, and at least the history of learning the deficiency of requirements as training data.
The estimation unit
The acquired attribute information of the applicant is applied to the first rule to obtain the estimation result of the information on the examination process.
An examination support device characterized in that the estimation result of the attribute information of the applicant and the estimation result of the information of the examination process are applied to the second rule to obtain the estimation result of deficiency of requirements. In the examination support device according to any one of claims 1 to 7.
The storage unit that generates the training data by accumulating information about the work of past application projects,
An examination support device including a learning unit that performs machine learning using the training data generated in the storage unit and generates the determination criteria. The acquisition step by the examination server to acquire the information of the application matter,
The estimation step by the support server to obtain the estimation result of the requirement deficiency by applying the acquired information of the application matter to the judgment criteria, and
The estimation result of the requirement deficiency is provided to the examination work of the application case, and the support step by the support server that supports the examination work is provided.
The criterion used in the estimation step is
An examination support method characterized in that it is the result of machine learning in which the attribute information of the applicant and the information on the examination process are input, and at least the history of learning the deficiency of requirements is used as training data. The examination support device according to any one of claims 1 to 8 provides information on the estimation result of the requirement deficiency to the examination work of the application matter or the applicant, and the examination work or the application of the applicant. A service provision method that provides services that support business.

Images