Intelligent auditing method and system for digital file and storage medium

Abstract

The invention provides an intelligent auditing method and system for a digital file and a storage medium, and the method comprises the steps: obtaining a case handling signal which comprises a case type; determining a first round of case handling questions corresponding to the case types, receiving case handling answers corresponding to the first round of case handling questions, and based on the case handling answers, dividing the settled cases in the database into strong-association settled cases and weak-association settled cases; obtaining a case handling question of the next round from the strong-association already-settled cases every time the strong-association already-settled cases or the weak-association already-settled cases are obtained, receiving case handling answers corresponding to the case handling questions, judging whether the case handling answers represent that case handling is completed or not, and if yes, generating a class case recommendation ending signal; and if not, re-dividing the settled cases in the database into strong-association settled cases and weak-association settled cases based on the case handling answers. The case handling efficiency can be improved, and the influence of experience of case handling personnel on case handling can be reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of case handling, in particular to an intelligent auditing method and system for digital files and a storage medium. BACKGROUND

[0002] The complexity of bad behavior and the unquantifiable nature of human factors determine that the handling of similar cases also needs the participation of case handling personnel to complete the handling of cases.

[0003] At the present stage, case handling personnel mainly rely on traditional experience mode to handle cases when auditing similar cases, that is, clues are sorted out by manually checking electronic or paper materials, and the nature of the case and the direction of investigation are judged according to personal case handling experience. In the evidence collection link, detection equipment and visits, witnesses and other means are used. In the case of case research and judgment, more reliance is placed on collective discussion of the special case group and guidance from the superiors. Although this method can complete the auditing and handling of similar cases, it takes a lot of time, making the auditing efficiency low and the experience of case handling personnel a major factor affecting the handling of cases. SUMMARY

[0004] In order to improve the auditing efficiency of case handling and reduce the influence of case handling personnel's experience on the handling of cases, the embodiments of the present application provide an intelligent auditing method and system for digital files and a storage medium.

[0005] In a first aspect, the embodiments provide an intelligent auditing method for digital files, which comprises: obtaining a case handling signal, wherein the case handling signal comprises a case type; determining a first round of case handling questions corresponding to the case type, receiving a case handling answer corresponding to the first round of case handling questions, and dividing the closed cases in the database into strongly associated closed cases and weakly associated closed cases based on the case handling answer; obtaining a next round of case handling questions from the strongly associated closed cases every time a strongly associated closed case or a weakly associated closed case is obtained, receiving a case handling answer corresponding to the case handling questions, and determining whether the case handling answer represents the completion of case handling, if so, generating a similar case recommendation end signal; if not, re-dividing the closed cases in the database into strongly associated closed cases and weakly associated closed cases based on the case handling answer.

[0006] In some of the embodiments, the case handling answer contains case information, case address and involved personnel, the strongly associated closed cases include first strongly associated closed cases, second strongly associated closed cases and third strongly associated closed cases, and the step of dividing the closed cases in the database into strongly associated closed cases and weakly associated closed cases based on the case handling answer comprises: determining the closed case containing only any one of the cause information, the case address and the involved personnel as the first strong correlation closed case; determining the closed case containing only any two of the cause information, the case address and the involved personnel as the second strong correlation closed case; determining the closed case containing all of the cause information, the case address and the involved personnel as the third strong correlation closed case; determining the closed cases other than the strong correlation closed cases as weak correlation cases.

[0007] In some embodiments, the obtaining the next round of case handling questions from the strong correlation closed cases each time a strong correlation closed case or a weak correlation closed case is obtained comprises: substituting all the obtained case handling answers into a prediction case analysis model to obtain a predicted case; substituting the recently obtained case handling answer into the predicted case to obtain case handling progress information corresponding to the recently obtained case handling answer; checking reference case handling questions of each strong correlation closed case at the next stage of the case handling progress information, and assigning corresponding scores to the corresponding reference case handling questions according to the level to which the strong correlation closed case belongs, wherein the first strong correlation closed case, the second strong correlation closed case and the third strong correlation closed case belong to different levels of cases, and each has a corresponding score; adding all the scores belonging to the same reference case handling question to obtain a total score corresponding to the reference case handling question, and determining the maximum score, and determining the question corresponding to the maximum score as the initial case handling question; determining whether a supplementary question is received, if not, the initial case handling question is the next round of case handling question; if yes, determining the initial case handling question and the supplementary question as the next round of case handling question.

[0008] In some embodiments, the method further comprises: obtaining a case handling answer, performing format verification on the case handling answer and the corresponding case handling question, and writing the case handling answer and the corresponding case handling question into the case file corresponding to the case handling signal after the verification is passed; if the verification is not passed, using a preset modification rule to modify the format of the case handling answer and / or the corresponding case handling question, and writing the modified case handling answer and / or the corresponding case handling question into the case file corresponding to the case handling signal; determining whether all case handling questions and case handling answers are written into the case file corresponding to the case handling signal, if yes, generating closing information for the case file.

[0009] In some embodiments, the receiving the case handling answer corresponding to the case handling question further comprises: obtaining the legal provisions involved in the predicted case, generating the legal provisions and pushing them to the case handling personnel.

[0010] In some embodiments, the method further comprises: For each predicted case, determining whether the predicted case deviates from the previous predicted case by more than a preset deviation, and if so, sending the predicted case to the case handling personnel, receiving feedback information from the case handling personnel, and modifying the predicted case based on the feedback information.

[0011] In some embodiments, the method further comprises: If the feedback information is non-blank information, the modified predicted case is fed back to the predicted case analysis model to adjust the predicted case analysis model.

[0012] In some embodiments, the method further comprises: receiving a question and answer request, and obtaining fourth strongly associated closed cases from the database corresponding to the question and answer request.

[0013] In a second aspect, the present embodiment provides an intelligent auditing system for digital files, which comprises an acquisition module, a processing module and a recommendation module; wherein, the acquisition module is configured to acquire a case handling signal, wherein the case handling signal comprises a case type; the processing module is configured to determine a first round of case handling questions corresponding to the case type; the recommendation module is configured to receive case handling answers corresponding to the first round of case handling questions, and divide the closed cases in the database into strongly associated closed cases and weakly associated closed cases based on the case handling answers; the processing module is further configured to obtain a next round of case handling questions from the strongly associated closed cases for each strongly associated closed case or weakly associated closed case, receive case handling answers corresponding to the case handling questions, and determine whether the case handling answers represent the completion of case handling, and if so, generate a case recommendation end signal; the recommendation module is configured to redivide the closed cases in the database into strongly associated closed cases and weakly associated closed cases based on the case handling answers if not.

[0014] In a third aspect, the present embodiment provides a computer readable storage medium having a computer program stored thereon, which can be run on a processor, and the computer program is executed by the processor to implement the intelligent auditing method for digital files according to the first aspect.

[0015] By adopting the method, the application first acquires a case handling signal, wherein the case handling signal comprises a case type. Then a first round of case handling questions corresponding to the case type is determined, a case handling answer corresponding to the first round of case handling questions is received, and the closed cases in the database are divided into strong correlation closed cases and weak correlation closed cases based on the case handling answer. Then, every time a strong correlation closed case or a weak correlation closed case is obtained, a next round of case handling questions is obtained from the strong correlation closed cases, a case handling answer corresponding to the case handling questions is received, and it is judged whether the case handling answer represents completion of case handling. If yes, a case recommendation end signal is generated; if no, the closed cases in the database are re-divided into strong correlation closed cases and weak correlation closed cases based on the case handling answer. In this way, case audit recommendation is not only performed once in the process of case handling, and case audit recommendation is not only performed at certain fixed stages, but also adjusted in real time according to the obtained information, so that the recommended cases can be more suitable for the case, and the promotion of the cases is automatically analyzed and processed by the model, thereby improving the case handling efficiency and reducing the influence of the experience of case handling personnel on case handling. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 FIG. 1 is a method block diagram of an intelligent audit method for digital files provided by an embodiment of the application.

[0017] Figure 2 FIG. 4 is a method block diagram of dividing the closed cases in the database into strong correlation closed cases and weak correlation closed cases based on the case handling answer provided by an embodiment of the application.

[0018] Figure 3 FIG. 6 is a method block diagram of obtaining a next round of case handling questions from the strong correlation closed cases every time a strong correlation closed case or a weak correlation closed case is obtained provided by an embodiment of the application.

[0019] Figure 4 FIG. 8 is a connection schematic diagram of an intelligent audit system for digital files provided by an embodiment of the application. DETAILED DESCRIPTION

[0020] To more clearly understand the purpose, technical solutions and advantages of the application, the application is described and explained below in conjunction with the drawings and embodiments. However, it should be understood by those skilled in the art that the application can be implemented without these details. It is obvious for those skilled in the art that various changes can be made to the embodiments disclosed in the application, and the general principles defined in the application can be applied to other embodiments and application scenarios without deviating from the principles and scope of the application. Therefore, the application is not limited to the shown embodiments, but conforms to the broadest scope claimed in the application.

[0021] The embodiments of the present application are further described in detail below with reference to the accompanying drawings.

[0022] In the present application, the case recommendation is an important associated means in the case handling of the case audit, and the case recommendation is realized by analyzing the case content, and the case handling personnel can improve the processing efficiency and processing logic of the similar cases based on the recommended cases.

[0023] Figure 1 A kind of intelligent audit method for digital files provided by the embodiment of the present application is shown in the figure. Figure 1 As shown in the figure, a kind of intelligent audit method for digital files includes the following steps: Step S100, obtaining case handling signal, wherein the case handling signal includes case handling type.

[0024] The embodiment is described from the central processing end. The above-mentioned case handling signal refers to the signal that needs to carry out case handling work. The case handling signal is sent by the case handling personnel to the central processing end when there is a need to carry out case handling, so that the central processing end obtains the case handling signal. Specifically, the case handling personnel receiving the case request of the public can indicate that there is a need to carry out case handling, and the case handling personnel can also indicate that there is a need to carry out case handling when they find that there is a case during patrol. Here, how to determine whether there is a need to carry out case handling is not limited further. The case handling signal includes case handling type, that is, which kind of case the case handling personnel needs to handle belongs to mild fighting case type, moderate fighting case type, severe fighting case type, theft case type, traffic accident case type, domestic violence case type, dispute case type, right case type, etc. The case handling personnel contains the case type corresponding to the case in the case handling signal sent to the central processing end.

[0025] Step S200, determine the first round of case handling problems corresponding to the case type, receive the case handling answers corresponding to the first round of case handling problems, and divide the closed cases in the database into strong correlation closed cases and weak correlation closed cases based on the case handling answers.

[0026] The database at the central processing end stores the basic questions corresponding to each case type, that is, when the case handling personnel starts to handle the case, the questions involved in the basic situation of the case need to be understood, that is, the first round of case handling questions need to be asked in handling the case. Therefore, by substituting the case type to be handled into the database, the basic questions corresponding to the case type are found, and the first round of case handling questions corresponding to the case type is determined. Then the case handling personnel obtains the first round of case handling questions, and takes corresponding actions according to the first round of case handling questions, thereby obtaining the case handling answers corresponding to the first round of case handling questions, and sends the obtained case handling answers to the central management end, so that the central management end receives the case handling answers corresponding to the first round of case handling questions. If the corresponding action of the first round of case handling questions includes asking questions to the person involved, the person involved can use the central processing end to directly obtain the answers of the person involved by communicating with the voice and image equipment deployed in the case handling room. If the person involved is not in the case handling room, the case handling personnel needs to ask questions to the person involved and record the answers of the person involved, and then upload the recorded answers to the central processing end, so that the central processing end receives the answers of the person involved.

[0027] If the corresponding action of the first round of case handling questions includes obtaining the video of a certain place, the relevant answers can be obtained by checking the equipment at the place and communicating with the central management end through the equipment, so that the central processing end receives the relevant answers. In this way, after obtaining the case handling questions, the corresponding case handling answers are automatically obtained by using the associated equipment as much as possible, which can reduce the number of intermediate transmissions and improve the accuracy of the obtained case handling answers, providing more accurate information reference for subsequent similar case recommendation.

[0028] After obtaining the case handling answers of this round, the closed cases in the database are divided into strong correlation closed cases and weak correlation closed cases based on the case handling answers. The strong correlation closed cases represent cases with strong correlation with the case, which are the similar cases to be recommended in this round. The weak correlation closed cases represent cases with low correlation with the case, which are not recommended in this round.

[0029] The case handling answers corresponding to the first round of case handling questions include case information, case address and person involved. The strong correlation closed cases include first strong correlation closed cases, second strong correlation closed cases and third strong correlation closed cases. The first strong correlation closed cases, the second strong correlation closed cases and the third strong correlation closed cases belong to different levels of cases, that is, different levels of cases indicate different degrees of similarity with the case. Among all the strong correlation closed cases, the degrees of similarity with the case in order from low to high are the first strong correlation closed cases, the second strong correlation closed cases and the third strong correlation closed cases. Figure 2The method block diagram for dividing the closed cases in the database into strong correlation closed cases and weak correlation closed cases based on the case handling answer is provided by the embodiment of the present application. As shown in Figure 2 The method for dividing the closed cases in the database into strong correlation closed cases and weak correlation closed cases based on the case handling answer includes the following steps: Step S201, reading the information of each closed case in the database, and determining the first order case containing only any one of the case information, case address and involved personnel as the first strong correlation closed case.

[0030] Step S202, determining the closed case containing only any two of the case information, case address and involved personnel as the second strong correlation closed case.

[0031] Step S203, determining the closed case containing the case information, case address and involved personnel as the third strong correlation closed case.

[0032] Step S204, determining the other cases in the closed cases except the strong correlation closed cases as the weak correlation cases.

[0033] The file header of each file of the closed case stored in the database is marked with several information which can indicate the key information and basic information of the case. The information of each closed case in the database can be obtained by checking the file header of each closed case. Then each information is compared with the case information, case address and involved personnel in the case handling answer to obtain the number of the information containing the case information, case address and involved personnel. Then the closed case corresponding to the information with the number one is determined as the first strong correlation closed case, the closed case corresponding to the information with the number two is determined as the second strong correlation closed case, the closed case corresponding to the information with the number three is determined as the third strong correlation closed case, and the closed case corresponding to the information with the number zero is determined as the weak correlation closed case. In this way, after obtaining the basic information of the case, the limited information comparison is used to quickly recommend the closed cases which may have strong correlation with the case from all the closed cases in the database, i.e. the similar case recommendation, which provides guidance for the subsequent further case handling, and can indirectly improve the case handling efficiency and reduce the influence of the experience of the case handling personnel on the case handling.

[0034] Step S300, obtaining the next round of case handling question from the strong correlation closed cases every time the strong correlation closed case or the weak correlation closed case is obtained, receiving the case handling answer corresponding to the case handling question, and judging whether the case handling answer represents the completion of the case handling. If yes, a similar case recommendation end signal is generated.

[0035] Step S400, if no, the closed cases in the database are re-divided into strong correlation closed cases and weak correlation closed cases based on the case handling answer.

[0036] The division of the closed cases in the database into strong correlation closed cases and weak correlation closed cases is to indicate that the case handling process is completed once the case recommendation is completed. Since the case handling is not completed in one round of questions and answers, multiple rounds of questions and answers are required, that is, the next round of case handling questions are obtained from the strong correlation closed cases each time a strong correlation closed case or a weak correlation closed case is obtained. Specifically, after obtaining the case handling answer corresponding to the first round of case handling questions and performing case recommendation, the next round of case handling questions need to be determined. Figure 3 is a method block diagram provided by an embodiment of the present application for obtaining the next round of case handling questions from the strong correlation closed cases each time a strong correlation closed case or a weak correlation closed case is obtained. As shown in Figure 3 each time a strong correlation closed case or a weak correlation closed case is obtained, the next round of case handling questions are obtained from the strong correlation closed cases, which includes the following steps: Step S301, substituting all obtained case handling answers into the pre-case analysis model to obtain the predicted case.

[0037] Step S302, substituting the recently obtained case handling answer into the predicted case to obtain the recently obtained case handling progress information.

[0038] Step S303, checking the reference case handling question of each strong correlation closed case at the next stage of the case handling progress information, and according to the grade to which the strong correlation closed case belongs, assigning the corresponding reference case handling question with a corresponding score, wherein the first strong correlation closed case, the second strong correlation closed case and the third strong correlation closed case belong to different grades of cases, and each has a corresponding score.

[0039] Step S304, adding all scores belonging to the same reference case handling question to obtain the total score corresponding to the reference case handling question, and determining the maximum score, and determining the problem corresponding to the maximum score as the initial case handling question.

[0040] Step S305, determining whether a supplementary question is received, if not, the initial case handling question is the next round of case handling question.

[0041] Step S306, if received, the initial case handling question and the supplementary question are determined as the next round of case handling question.

[0042] The above-mentioned pre-case analysis model is specifically a model trained by all the closed cases of the case type in the database, so that the pre-case analysis model is more in line with the case analysis of the case. The central processing end will input all the case handling answers of the case into the pre-case analysis model after receiving a round of case handling answers, so that the pre-case analysis model processes and outputs a predicted case at the output end, that is, the pre-case analysis model predicts a probable case of the case based on all the currently obtained case handling answers.

[0043] Then, the recently obtained case handling answers of this round are input into the pre-case, and compared with the information in the pre-case to obtain which contents of the pre-case are disclosed by the recently obtained case handling answers of this round, so as to obtain the case handling progress information corresponding to the recently obtained case handling answers of this round according to the sorting positions of the contents in the pre-case.

[0044] Then, the questions of each strongly correlated closed case at the next stage of the case handling progress information are checked to obtain reference case handling questions, that is, the reference case handling questions corresponding to each strongly correlated closed case are obtained, and the score corresponding to each reference case handling question is the score corresponding to the strongly correlated closed case. Each level of strongly correlated case corresponds to a corresponding score, and the corresponding scores from small to large are the first strongly correlated closed case, the second strongly correlated closed case, and the third strongly correlated closed case. The scores corresponding to different levels of strongly correlated cases are different. Then, the scores corresponding to the same question in all reference case handling questions are added to obtain the total score corresponding to the question. If some questions in all reference case handling questions do not belong to the same question, the total score of the question is the score corresponding to the question. Thus, the total score corresponding to each reference case handling question is obtained, and by comparing all the total scores, the maximum score can be determined, and the reference case handling question corresponding to the maximum score is the initial case handling question. The initial case handling question is represented as a question that needs to be proposed by the central processing end in the next round.

[0045] However, the case handling personnel may also have their own ideas during case handling, so there are some questions. In order to more comprehensively and accurately handle the case, the central processing end can receive the supplementary questions sent by the case handling personnel when the case handling personnel has questions, that is, when the supplementary questions are received, the supplementary questions and the above-mentioned initial case handling question are determined as the case handling questions in the next round. When no supplementary questions are received, the above-mentioned initial case handling question is determined as the case handling question in the next round. In this way, when the central processing end automatically integrates the case handling questions needed in the next round based on the promoted similar cases, the case handling personnel's supplementary questions for the case are also considered, so that each case handling question is more comprehensive and accurate.

[0046] The person handling the case gets the case handling question again, and then takes corresponding action according to the case handling question, so as to get the case handling answer corresponding to the case handling question, and then sends the obtained case handling answer to the central management terminal, so that the central management terminal receives the case handling answer corresponding to the case handling question.

[0047] After obtaining the case handling answer, the case handling answer is substituted into the obtained pre-case situation to obtain case handling progress information corresponding to the case handling answer, that is, to determine whether the sorting position of all obtained information in the pre-case situation is the last position. If it is the last position, it indicates that the case handling answer represents completed case handling. If it is not the last position, it indicates that the case handling answer does not represent completed case handling.

[0048] If it is judged that the case handling answer represents completed case handling, the central processing terminal generates a case recommendation end signal to stop the subsequent case recommendation of the case. If it is judged that the case handling answer does not represent completed case handling, the central processing terminal extracts a plurality of core information in the case handling answer, and then uses the above-mentioned method to re-divide the closed cases in the database into strong correlation closed cases and weak correlation closed cases using the case handling answer, so as to execute the above-mentioned step S300. In this way, case recommendation is not only performed once in the process of handling the case, and case recommendation is not only performed at some fixed stage, but also adjusted in real time according to the obtained information, so that the recommended cases can be more suitable for the case, and the promotion of the cases is automatically analyzed and processed by the model, thereby improving the efficiency of handling the case and reducing the influence of the experience of the case handling personnel on the case handling.

[0049] When the closed cases in the database are re-divided into strong correlation closed cases and weak correlation closed cases using the case handling answer, the strong correlation closed cases and the weak correlation closed cases of the current round can be determined from the strong correlation closed cases already divided in the last round, and the strong correlation closed cases and the weak correlation closed cases of the current round can be determined from the weak correlation closed cases already divided in the last round. In this way, the strong correlation closed cases of the current round can be obtained more quickly, thereby indirectly improving the efficiency of handling the case.

[0050] Preferably, the central processing terminal performs format verification on the case handling answer and the corresponding case handling question every time a case handling answer is obtained, and writes the case handling answer and the corresponding case handling question into the case file corresponding to the case handling information after the verification is passed. If the verification fails, the case handling answer and / or the corresponding case handling question are modified using a preset modification rule, and the modified case handling answer and / or the corresponding case handling question are written into the case file corresponding to the case handling information. It is judged whether all case handling questions and case handling answers are written into the case file corresponding to the case handling information. If yes, the case file is generated into a closed case information.

[0051] Specifically, the central management end will use the associated file assistant program to format verify the case handling answer and the corresponding case handling question after obtaining the case handling answer, and automatically adjust the corresponding format when there are format problems such as typos, incorrect formats, and incorrect reference relationships, to obtain the verified information, and then write the verified information into the case file corresponding to the case handling signal. When there are no format problems such as typos, incorrect formats, and incorrect reference relationships, the case handling answer and the corresponding case handling question are taken as the verified information, and the information is written into the case file corresponding to the case handling signal. Each case handling signal corresponds to a blank file in the database, and the central processing end writes the verified information into the blank file after obtaining the information, so that all case handling questions and case handling answers of the case after verification are written into the case file corresponding to the case handling signal, and the file assistant program is used to arrange and extract all the written information to obtain the case closing information, which is stored in the database and becomes a closed case. In this way, the information is arranged gradually during the process of obtaining the case handling question and the corresponding case handling answer, so that the corresponding file can be arranged more quickly when the case is completed, the time spent on arranging the file after the case is clear is reduced, and the completion efficiency of the case is improved.

[0052] Preferably, the central processing end also obtains the legal provisions involved in the predicted case after receiving the case handling answer corresponding to the case handling question, generates the legal provisions, and pushes them to the case handling personnel.

[0053] The central processing end sends the received predicted case to the associated file assistant program, which reads the predicted case and obtains the legal provisions involved in the predicted case by comparing with the stored legal provisions, and pushes them to the case handling personnel in the form of voice or text, so that the case handling personnel can decide whether to inform the relevant personnel based on the actual situation. This can assist the case handling personnel in handling the case.

[0054] Preferably, the central processing end judges whether the deviation between the predicted case and the case of the previous predicted case is greater than a preset deviation after obtaining each predicted case, and if it is greater, sends the predicted case to the case handling personnel, receives the feedback information of the case handling personnel, and modifies the predicted case based on the feedback information.

[0055] Specifically, the case deviation is specifically the number of different information between the predicted case and the previous predicted case, and the preset deviation is the number of allowed different information when the cases are the same by default. By comparing the number of different information between the predicted case and the previous predicted case with the number of allowed different information, if the number of different information between the predicted case and the previous predicted case is less than the number of allowed different information, it indicates that the case deviation between the predicted case and the previous predicted case located above the predicted case is not greater than the preset deviation, that is, the obtained predicted case is accurate, and at this time, the obtained predicted case is taken as the predicted case of this time. If the number of different information between the predicted case and the previous predicted case is not less than the number of allowed different information, it indicates that the case deviation between the predicted case and the previous predicted case located above the predicted case is greater than the preset deviation, that is, the obtained predicted case is inaccurate, and at this time, the obtained predicted case is sent to the case handling personnel. When it is determined to be accurate, the case handling personnel will send a blank feedback information. When it is determined to be inaccurate, the case handling personnel will send a feedback information containing an accurate predicted case, that is, the feedback information is non-blank information. The central processing end will modify the predicted case according to the received feedback information, so as to obtain an accurate predicted case. In this way, by judging the accuracy of the predicted case, the accuracy of the subsequent determination of the case handling problem can be ensured, and then the accuracy of the subsequent recommendation of similar cases can be ensured. After all, the accuracy of the predicted case will determine the accuracy of the subsequent promotion of similar cases.

[0056] When the feedback information is non-blank information, the modified predicted case is fed back to the predicted case analysis model to adjust the predicted case analysis model. When the accuracy of the predicted case obtained by the predicted case analysis model is reduced, the predicted case analysis model is adjusted in time, which can improve the accuracy of the subsequent obtained predicted case, and further improve the more accurate recommendation of similar cases.

[0057] Preferably, the central processing end also receives a question and answer request, and divides the case handled into a fourth strong correlation case handled from the database corresponding to the question and answer request. The question and answer request refers to the information needed by the case handling personnel in the case handling process. The information has a strong correlation with the case handling, and therefore, the case handled corresponding to the question and answer request is divided into the fourth strong correlation case handled, so that the proportion of the fourth strong correlation case handled to the case recommendation of the case is greater, thereby increasing the effective guidance of the case handling personnel to the case and better helping the handling of the case.

[0058] Figure 4 It is a connection diagram of an intelligent auditing system for digital files provided by the embodiment of the application. As shown in Figure 4 The intelligent auditing system for digital files includes an acquisition module, a processing module and a recommendation module.

[0059] The obtaining module is configured to obtain a case handling signal, wherein the case handling signal comprises a case type. The processing module is configured to determine a first round of case handling questions corresponding to the case type. The recommendation module is configured to receive a case handling answer corresponding to the first round of case handling questions, and divide the closed cases in the database into strongly-correlated closed cases and weakly-correlated closed cases based on the case handling answer. The processing module is further configured to obtain a next round of case handling questions from the strongly-correlated closed cases each time a strongly-correlated closed case or a weakly-correlated closed case is obtained, receive a case handling answer corresponding to the case handling questions, and determine whether the case handling answer represents completion of the case handling. If yes, a case recommendation end signal is generated. If no, the closed cases in the database are re-divided into strongly-correlated closed cases and weakly-correlated closed cases based on the case handling answer.

[0060] The other functions performed by the above-described obtaining module, processing module and recommendation module, and the technical details of each function are the same as or similar to the corresponding features of the above-described intelligent auditing method for digital files, and thus will not be described again here.

[0061] The embodiments of the present application also provide a computer storage medium having a computer program stored thereon, which, when executed on a computer, enables the computer to perform the steps of the above-described intelligent auditing method for digital files.

[0062] It should be understood that, although each step in the flowchart of the accompanying drawings is shown in sequence according to the direction of the arrow, these steps are not necessarily executed in sequence according to the direction of the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and they can be executed in other sequences.

[0063] The above only describes some embodiments of the present application, and it should be noted that, for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which should also be considered as falling within the scope of protection of the present application.

Claims

1. An intelligent audit method for digital files, characterized in that: The method comprises: Acquiring a case handling signal, wherein the case handling signal includes a case type; determining a first round of case handling questions corresponding to the case type, receiving case handling answers corresponding to the first round of case handling questions, and classifying closed cases in the database into strongly associated closed cases and weakly associated closed cases based on the case handling answers; Each time a strongly related closed case or a weakly related closed case is obtained, the next round of case handling questions is obtained from the strongly related closed case, the case handling answers corresponding to the case handling questions are received, and it is determined whether the case handling answers indicate that the case handling is completed. If so, a similar case recommendation end signal is generated; If not, the closed cases in the database are reclassified into strongly-related closed cases and weakly-related closed cases based on the case handling answers.

2. The method according to claim 1, characterized in that The case handling answer includes case cause information, case address, and persons involved in the case. The strongly associated closed cases include a first strongly associated closed case, a second strongly associated closed case, and a third strongly associated closed case. The classification of closed cases in the database into strongly associated closed cases and weakly associated closed cases based on the case handling answer includes: Read the information of each closed case in the database, and determine the closed case that only contains any one of the case cause information, case address, and involved persons as the first strongly associated closed case; The closed cases that only contain any two of the case cause information, case address, and persons involved are determined as the second most strongly correlated closed cases; The closed cases that all contain the case cause information, case address, and persons involved are determined as the third most strongly correlated closed cases; The other cases among the closed cases except the strongly related closed cases are determined as weakly related cases.

3. The method according to claim 1, characterized in that Each time a strongly related closed case or a weakly related closed case is obtained, the next round of case handling questions are obtained from the strongly related closed case, including: Substitute all the obtained case handling answers into the predicted case analysis model to obtain the predicted case situation; Substituting the most recently obtained case handling answer into the predicted case situation to obtain case handling progress information corresponding to the most recently obtained case handling answer; Reviewing the reference case handling questions for each strongly associated closed case at the next stage of the case handling progress information, and assigning corresponding scores to the corresponding reference case handling questions according to the level of the strongly associated closed case, wherein the first strongly associated closed case, the second strongly associated closed case, and the third strongly associated closed case belong to different levels of cases and each has a corresponding score; Adding up all scores belonging to the same reference case handling question to obtain a total score corresponding to the reference case handling question, determining the maximum score, and determining the question corresponding to the maximum score as the initial case handling question; Determine whether supplementary questions have been received. If not, the initial case handling questions will be used as the case handling questions for the next round; If received, the initial case handling questions and the supplementary questions will be determined as the case handling questions for the next round.

4. The method according to claim 1, wherein The method further comprises: Each time a case handling answer is obtained, the format of the case handling answer and the corresponding case handling question is verified, and after passing the verification, they are written together into the case file corresponding to the case handling signal; If the verification fails, the format of the case handling answer and / or the corresponding case handling question is modified using the preset modification rules, and the modified case handling answer and / or the corresponding case handling question is written into the case file corresponding to the case handling signal; Determine whether all case handling questions and answers are written into the case file corresponding to the case handling signal. If so, generate case closing information for the case file.

5. The method according to claim 3, characterized in that After receiving the case handling answer corresponding to the case handling question, the method further includes: Obtain the legal provisions involved in the predicted case, generate the legal provisions and push them to the case handlers.

6. The method according to claim 3, characterized in that The method further comprises: Every time a predicted case is obtained, it is determined whether the case deviation between the predicted case and the previous predicted case is greater than the preset deviation. If it is, the predicted case is sent to the case handler, feedback information from the case handler is received, and the predicted case is modified based on the feedback information.

7. The method according to claim 6, characterized in that The method further comprises: If the feedback information is non-blank information, the modified predicted case situation is fed back to the predicted case situation analysis model to adjust the predicted case situation analysis model.

8. The method according to claim 1, characterized in that The method further comprises: A question-and-answer request is received, and closed cases corresponding to the question-and-answer request are obtained from a database and classified as the fourth most strongly associated closed cases.

9. An intelligent audit system for digital files, characterized by: The system includes an acquisition module, a processing module and a recommendation module; wherein, The acquisition module is used to acquire a case handling signal, wherein the case handling signal includes a case type; The processing module is used to determine the first round of case handling issues corresponding to the case type; The recommendation module is configured to receive the case handling answers corresponding to the case handling questions in the first round, and classify the closed cases in the database into strongly associated closed cases and weakly associated closed cases based on the case handling answers; The processing module is further configured to obtain, each time a strongly associated closed case or a weakly associated closed case is obtained, a next round of case handling questions from the strongly associated closed cases, receive a case handling answer corresponding to the case handling question, and determine whether the case handling answer indicates that the case handling is completed. If so, generate a similar case recommendation end signal; The recommendation module is used to reclassify the closed cases in the database into strongly associated closed cases and weakly associated closed cases based on the case handling answers.

10. A computer-readable storage medium storing a computer program that can be run on a processor, characterized in that: When the computer program is executed by the processor, it implements an intelligent audit method for digital files as described in any one of claims 1 to 8.

Images