Community dispute rapid mediation system based on self-learning

Abstract

This invention provides a self-learning-based rapid mediation system for community disputes. The system includes a database module, an account login module, an input module, and a processing module. The database module stores solutions to community disputes. The account login module is used for user account login. The input module is used for user input of dispute information. The processing module includes a matching unit and a self-learning unit. The matching unit identifies and processes the dispute information, categorizes it, and matches the categorized information with pre-stored solutions. This invention can match corresponding solutions based on the nature of the community dispute and can self-learn to match solutions for new dispute information, improving the intelligence of community dispute resolution and addressing the problems of limited solutions and heavy workload in existing community dispute resolution methods.

Description

Technical Field

[0001] The present invention relates to the technical field of community management systems, and in particular to a community dispute rapid mediation system based on self-learning. Background Art

[0002] Community management refers to a series of self-management or administrative management activities carried out by various institutions, groups or organizations within a certain community in order to maintain the normal order of the community, promote the development and prosperity of the community, and meet the specific needs of community residents such as material and cultural activities. Community management is the self-management of various public affairs and public welfare undertakings of the community by community functional departments, community units and community residents under the guidance of the government.

[0003] In the existing community management, since there are a large number of residents living in the community and the residents live very close to each other and live in the same community environment, there are many problems in the community, such as littering, dripping air-conditioning outdoor units, sewer blockages, water accumulation in patios, sewage overflows, and occupation of common spaces. Many problems have standardized solutions. If every community dispute requires community staff to solve them one by one, then the workload of community management will be huge. The existing management system for community disputes is relatively simple in function and it is difficult to solve the increasing number of community disputes. Summary of the invention

[0004] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a community dispute rapid mediation system based on self-learning, which can match corresponding solutions according to the problems of community disputes, and can self-learn and match solutions for new dispute information, thereby improving the intelligence of problem solving in community disputes, so as to solve the problems that the existing methods of solving community disputes are relatively single and the workload of community management is large.

[0005] In order to achieve the above-mentioned object, the present invention is implemented through the following technical solutions: a community dispute rapid mediation system based on self-learning, the mediation system includes a database module, an account login module, an input module and a processing module;

[0006] The database module stores solutions to community disputes;

[0007] The account login module is used for users to log in to their accounts;

[0008] The input module is used for users to input dispute information;

[0009] The processing module includes a matching unit and a self-learning unit, wherein the matching unit is used to identify and process dispute information, classify the dispute information, match the classified dispute information with pre-stored solutions, and output the matched solutions to the user;

[0010] The self-learning unit is used to identify new dispute information that cannot be matched with a solution, and to rank and recommend approximate solutions, and to summarize the types according to the user's selection, and to obtain and store the types of solutions to the new dispute information.

[0011] Furthermore, the input module includes a pre-classification unit, and the pre-classification unit is used to classify and mark the dispute information;

[0012] The pre-classification unit is configured with a pre-classification strategy, which includes: marking the pre-stored dispute information one by one, marking them as J1 to Jn in sequence, where J1 represents the number of the first dispute information, Jn represents the number of the nth dispute information, and there are n pre-stored dispute information; then hierarchically classifying the n dispute information into first-level dispute types, second-level dispute types and third-level dispute types, the first-level dispute type is set as the upper-level summary type of the second-level dispute type, the second-level dispute type is set as the upper-level summary type of the third-level dispute type, and the third-level dispute types correspond to the n dispute information respectively.

[0013] Furthermore, the database module is configured with a solution marking strategy, which includes: marking the corresponding solution for each pre-stored dispute information according to its label, marking them in sequence as F1 to Fn, F1 represents the label of the solution corresponding to the first dispute information, Fn represents the label of the solution corresponding to the nth dispute information, and the n solutions correspond one-to-one to the n dispute information; then classifying the n solutions according to the hierarchy of the dispute information, and dividing them into first-level solution type, second-level solution type and third-level solution type, the first-level solution type is set to the upper-level summary type of the second-level solution type, the second-level solution type is set to the upper-level summary type of the third-level solution type, and the third-level solution types correspond to the n solutions respectively.

[0014] Furthermore, the input module further includes a custom input unit, and the custom input unit is used for a user to customize the type of dispute information input;

[0015] The custom input unit is configured with a custom input strategy, the custom input strategy includes: dividing the user's custom input information into a number of custom input types, the custom input types including a text input type, a picture input type, a video input type and a voice input type;

[0016] The text input type is configured with a text input box, and the text input box is used for the user to input text information of the dispute information;

[0017] The picture input type is configured with a picture upload box, and the picture upload box is used for users to upload picture information of dispute information;

[0018] The video input type is configured with a video upload box, and the video upload box is used for users to upload video information of dispute information;

[0019] The voice input type is configured with a voice input box, and the voice input box is used for users to upload voice information of dispute information.

[0020] Furthermore, the matching unit includes an identification subunit, and the identification subunit is configured with an identification strategy; the identification strategy includes: identifying the label corresponding to the information selected by the user according to the label of the dispute information;

[0021] The user-defined dispute information input by the user is identified in sequence; the text information of the text input type is semantically extracted, and a piece of dispute information after a label is matched according to the semantics, and the unrecognized text information is marked as unrecognized text information;

[0022] Identify the image information of the image input type, identify the occurrence scene and the status of the characters in the image information, match a labeled dispute information according to the occurrence scene and the status of the characters, and mark the unrecognized image information as unrecognized image information;

[0023] Identify the video information of the video input type, identify the occurrence scene and character status of the video information, match a labeled dispute information according to the occurrence scene and character status, and mark the unidentified video information as unidentified video information;

[0024] Recognize the voice information of voice input type, convert the voice information into text information, extract the semantics of the text information, match a labeled dispute information according to the semantics, and mark the unrecognized voice information as unrecognized voice information.

[0025] Furthermore, the matching unit further comprises a matching subunit, wherein the matching subunit is configured with a matching strategy, wherein the matching strategy comprises: selecting a solution corresponding to the number from a database module according to the number of the dispute information after the numbering, and sending the solution to the user;

[0026] Manual service is provided for information matching of unrecognized custom input.

[0027] Furthermore, the self-learning unit includes an induction subunit, and the induction subunit is configured with an induction strategy, and the induction strategy includes: after sending the labeled solution to the user, setting an option, the option is set to solution matching and solution mismatching; when the solution mismatches, selecting the three closest new solutions according to the type of dispute information for the user to choose in turn; when the three solutions still do not match, recommending the solutions in turn according to the first-level solution type, the second-level solution type and the third-level solution type;

[0028] Record the number of the solution finally selected by the user in sequence, and record the number of times several solutions are finally selected for the same dispute information, and calculate the correction value according to the correction algorithm;

[0029] When the correction value is greater than or equal to the first correction threshold, the solution with the largest correction value replaces the pre-matched solution, the pre-matched solution is marked as a waiting labeling solution, and its label is repeatedly marked to the final selected solution.

[0030] Furthermore, the correction algorithm is configured as: Pjz = k1 × (Tc) -Xh , where Pjz is the correction value, k1 is the correction coefficient, and k1 is greater than 0, TC is the number of times the solution appears, and Xh is the serial number of the solution sorted from large to small according to the number of times it appears.

[0031] Furthermore, the self-learning unit also includes a sorting subunit, which is configured with a sorting strategy. The sorting strategy includes: sorting and recommending solutions for unrecognized custom information according to relevance values ​​from large to small, the relevance values ​​are calculated based on a relevance value algorithm, and the sorting scheme is sent to manual verification. If the manually verified solution does not correspond to any of the top three solutions, the weight in the relevance value algorithm is adjusted.

[0032] Further, the correlation value algorithm is configured as: Pxg=a1×Yyb+a2×Tpb+a3×Spb, where Pxg is the correlation value, Yyb is the semantic matching degree of text information, Tpb is the picture matching degree of picture information, Spb is the video matching degree of video information, a1 is the first adjustment weight, a2 is the second adjustment weight, a3 is the third adjustment weight, and a1, a2 and a3 are respectively greater than 0.

[0033] Beneficial effects of the present invention: The present invention identifies and classifies dispute information, and matches the stored solutions with the classified dispute information one by one, so that the corresponding solutions can be outputted immediately when the user inputs the dispute information, thereby improving the efficiency of dispute handling;

[0034] The present invention identifies new dispute information that cannot be matched with a solution, ranks and recommends approximate solutions, summarizes the types according to the user's selection, and stores the types of solutions to the new dispute information. This method can self-learn and update the solutions to new disputes, thereby improving the comprehensiveness of the system's dispute resolution and the intelligence of community dispute resolution. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0036] Figure 1 It is a system principle block diagram of the present invention;

[0037] Figure 2 This is a principle block diagram of the first embodiment.

[0038] In the figure, 1. Mediation system; 11. Account login module; 12. Input module; 121. Pre-classification unit; 122. Custom input unit; 13. Processing module; 131. Matching unit; 1311. Identification subunit; 1312. Matching subunit; 132. Self-learning unit; 1321. Summarization subunit; 1322. Sorting subunit; 14. Database module; 2. User. DETAILED DESCRIPTION

[0039] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0040] See also Figure 1 , a community dispute rapid mediation system 1 based on self-learning, the mediation system 1 includes a database module 14, an account login module 11, an input module 12 and a processing module 13;

[0041] The database module 14 stores solutions to community disputes;

[0042] The account login module 11 is used by user 2 to log in to an account;

[0043] The input module 12 is used for user 2 to input dispute information;

[0044] The processing module 13 includes a matching unit 131 and a self-learning unit 132. The matching unit 131 is used to identify and process dispute information, classify the dispute information, match the classified dispute information with pre-stored solutions, and output the matched solutions to the user 2;

[0045] The self-learning unit 132 is used to identify new dispute information that cannot be matched with a solution, and to rank and recommend approximate solutions, and to summarize the types according to the selection of the user 2, and to obtain and store the types of solutions to the new dispute information.

[0046] For example, see Figure 2 , the input module 12 includes a pre-classification unit 121, and the pre-classification unit 121 is used to classify and mark the dispute information;

[0047] The pre-classification unit 121 is configured with a pre-classification strategy, which includes: marking the pre-stored dispute information one by one, marking them as J1 to Jn in sequence, J1 represents the number of the first dispute information, Jn represents the number of the nth dispute information, and the pre-stored dispute information is n; then hierarchically classifying the n dispute information into first-level dispute types, second-level dispute types and third-level dispute types, the first-level dispute type is set as the upper summary type of the second-level dispute type, the second-level dispute type is set as the upper summary type of the third-level dispute type, and the third-level dispute types correspond to the n dispute information respectively.

[0048] For example, the first-level dispute types are divided into neighborhood disputes, property service disputes, community management disputes, management system disputes, and birth management disputes. Among them, the second-level dispute types of property service disputes are divided into property standardized solution disputes and property manual solution disputes. For example, property standardized solution disputes are divided into garbage cleaning time range, road repairs, corridor renovations, etc.

[0049] The input module 12 further includes a custom input unit 122, and the custom input unit 122 is used for user 2 to customize the type of dispute information input;

[0050] The custom input unit 122 is configured with a custom input strategy, which includes: dividing the custom input information of the user 2 into a number of custom input types, the custom input types including a text input type, a picture input type, a video input type and a voice input type;

[0051] The text input type is configured with a text input box, and the text input box is used for user 2 to input text information of dispute information;

[0052] The picture input type is configured with a picture upload box, and the picture upload box is used for user 2 to upload picture information of dispute information;

[0053] The video input type is configured with a video upload box, and the video upload box is used for user 2 to upload video information of dispute information;

[0054] The voice input type is configured with a voice input box, and the voice input box is used for user 2 to upload voice information of dispute information.

[0055] The database module 14 stores solutions to community disputes; the database module 14 is configured with a solution marking strategy, which includes: marking the corresponding solution for each pre-stored dispute information according to its label, marking them in sequence as F1 to Fn, F1 represents the label of the solution corresponding to the first dispute information, Fn represents the label of the solution corresponding to the nth dispute information, and the n solutions correspond one-to-one to the n dispute information; then classifying the n solutions according to the hierarchy of the dispute information, and dividing them into first-level solution type, second-level solution type and third-level solution type, the first-level solution type is set to the upper-level summary type of the second-level solution type, the second-level solution type is set to the upper-level summary type of the third-level solution type, and the third-level solution types correspond to the n solutions respectively.

[0056] The matching unit 131 includes an identification subunit 1311 and a matching subunit 1312; the identification subunit 1311 is configured with an identification strategy; the identification strategy includes: identifying the label corresponding to the information selected by the user 2 according to the label of the dispute information;

[0057] The customized dispute information input by user 2 is recognized in sequence; semantic extraction is performed on the text information of the text input type, and a piece of dispute information after a label is matched according to the semantics, and the unrecognized text information is marked as unrecognized text information;

[0058] Identify the image information of the image input type, identify the occurrence scene and the status of the characters in the image information, match a labeled dispute information according to the occurrence scene and the status of the characters, and mark the unrecognized image information as unrecognized image information;

[0059] Identify the video information of the video input type, identify the occurrence scene and character status of the video information, match a labeled dispute information according to the occurrence scene and character status, and mark the unidentified video information as unidentified video information;

[0060] Recognize the voice information of voice input type, convert the voice information into text information, extract the semantics of the text information, match a labeled dispute information according to the semantics, and mark the unrecognized voice information as unrecognized voice information.

[0061] The matching subunit 1312 is configured with a matching strategy, which includes: selecting a solution corresponding to the number from the database module 14 according to the number of the dispute information after the numbering, and sending the solution to the user 2;

[0062] Manual service is provided for information matching of unrecognized custom input.

[0063] The self-learning unit 132 includes an induction subunit 1321 and a sorting subunit 1322. The induction subunit 1321 is configured with an induction strategy, which includes: after sending the numbered solution to the user 2, setting an option, the option is set to solution matching and solution mismatching; when the solution mismatches, selecting the three closest new solutions according to the type of dispute information for the user 2 to choose in turn; when the three solutions still do not match, recommending the solutions in turn according to the first-level solution type, the second-level solution type and the third-level solution type;

[0064] Record the number of the solution finally selected by user 2 in sequence, and record the number of times several solutions are finally selected for the same dispute information, and calculate the correction value according to the correction algorithm;

[0065] When the correction value is greater than or equal to the first correction threshold, the solution with the largest correction value replaces the pre-matched solution, the pre-matched solution is marked as a waiting labeling solution, and its label is repeatedly marked to the final selected solution.

[0066] The correction algorithm is configured as: Pjz = k1 × (Tc) -Xh , where Pjz is the correction value, k1 is the correction coefficient, and k1 is greater than 0, TC is the number of times the solution appears. The more the solution is selected, the closer it is to the correct solution, approximate solution or reference solution. Xh is the serial number of the solution sorted from large to small according to the number of times it appears. Sorting by the number of times the solution appears can express the correlation between the solution and the number of times it appears.

[0067] The sorting subunit 1322 is configured with a sorting strategy, which includes: sorting and recommending solutions for unrecognized custom information according to the relevance value from large to small, the relevance value is calculated based on the relevance value algorithm, and the sorting scheme is sent to manual verification. If the manually verified solution does not correspond to any of the top three solutions, the weight in the relevance value algorithm is adjusted.

[0068] The correlation value algorithm is configured as follows: Pxg=a1×Yyb+a2×Tpb+a3×Spb, where Pxg is the correlation value, Yyb is the semantic matching degree of text information, where text information also includes text information after voice information recognition, for text information that cannot be recognized, keywords can be extracted, and the semantic matching degree is obtained according to the degree of overlap of keywords, Tpb is the image matching degree of image information, and the image matching degree is obtained by obtaining similar information in the image, when the image information cannot obtain key information, an approximate matching scheme is adopted, and the image matching degree is assigned according to the similarity of the images, the higher the approximate degree, the higher the corresponding image matching degree, Spb is the video matching degree of video information, when the video information cannot be matched, an approximate matching scheme is also adopted, the video is matched with the closest video type, and the video matching degree is assigned according to the similarity, the higher the approximate degree, the higher the corresponding video matching degree, a1 is the first adjustment weight, a2 is the second adjustment weight, a3 is the third adjustment weight, and a1, a2 and a3 are respectively greater than 0.

[0069] Working principle: User 2 logs in to the system through the account login module 11. After logging in, he selects dispute information that is the same as his own needs according to the classification of the dispute information. The processing module 13 can match the corresponding solution for user 2. User 2 can also input according to his own needs. The input method can be text input, picture input, video input or voice input. The processing module 13 can identify the input information and match it to the corresponding dispute information, and then match the corresponding solution according to the dispute information. At the same time, the dispute information that cannot be identified can be matched with a solution after self-learning processing by the processing module 13, thereby improving the efficiency and intelligence of community dispute resolution.

[0070] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A community dispute rapid mediation system based on self-learning, characterized by: The mediation system (1) comprises a database module (14), an account login module (11), an input module (12) and a processing module (13); The database module (14) stores solutions to community disputes; The account login module (11) is used for the user (2) to log in to an account; The input module (12) is used for the user (2) to input dispute information; The input module (12) comprises a pre-classification unit (121), and the pre-classification unit (121) is used to classify and mark the dispute information; The pre-classification unit (121) is configured with a pre-classification strategy, the pre-classification strategy comprising: marking the pre-stored dispute information one by one, marking them sequentially as J1 to Jn, J1 represents the number of the first dispute information, Jn represents the number of the nth dispute information, and the pre-stored dispute information is n pieces; then classifying the n pieces of dispute information hierarchically, and respectively dividing them into first-level dispute types, second-level dispute types and third-level dispute types, the first-level dispute type is set as the upper-level summary type of the second-level dispute type, the second-level dispute type is set as the upper-level summary type of the third-level dispute type, and the third-level dispute types correspond to the n pieces of dispute information respectively; The processing module (13) includes a matching unit (131) and a self-learning unit (132), wherein the matching unit (131) is used to identify and process the dispute information, classify the dispute information, match the classified dispute information with a pre-stored solution, and output the matched solution to the user (2); The self-learning unit (132) is used to identify new dispute information that cannot be matched with a solution, and to rank and recommend approximate solutions, and to summarize the types according to the selection of the user (2), and to obtain the types of solutions for the new dispute information for storage; The self-learning unit (132) includes an induction subunit (1321), and the induction subunit (1321) is configured with an induction strategy, and the induction strategy includes: after sending the labeled solution to the user (2), setting an option, the option is set to solution matching and solution mismatching; when the solution mismatches, selecting the three closest new solutions according to the type of dispute information for the user (2) to select in sequence; when the three solutions still do not match, recommending the solutions in sequence according to the first-level solution type, the second-level solution type, and the third-level solution type; Record the number of the solution finally selected by the user (2) in sequence, and record the number of times several solutions are finally selected for the same dispute information, and calculate the correction value according to the correction algorithm; When the correction value is greater than or equal to the first correction threshold, the solution with the largest correction value replaces the pre-matched solution, the pre-matched solution is marked as a waiting labeling solution, and its label is repeatedly marked to the final selected solution; The database module (14) is provided with a solution marking strategy, the solution marking strategy comprising: marking the corresponding solution for each pre-stored dispute information according to its label, marking them in sequence as F1 to Fn, F1 represents the label of the solution corresponding to the first dispute information, Fn represents the label of the solution corresponding to the nth dispute information, and the n solutions correspond to the n dispute information one by one; then classifying the n solutions according to the level of the dispute information, and respectively dividing them into a first-level solution type, a second-level solution type and a third-level solution type, the first-level solution type is set as the upper-level summary type of the second-level solution type, the second-level solution type is set as the upper-level summary type of the third-level solution type, and the third-level solution type corresponds to the n solutions respectively; The correction algorithm is configured as: Pjz = k1 × (Tc) -Xh , where Pjz is the correction value, k1 is the correction coefficient, and k1 is greater than 0, TC is the number of times the solution appears, and Xh is the sequence number of the solution sorted from large to small according to the number of times it appears; The self-learning unit (132) further comprises a sorting subunit (1322), wherein the sorting subunit (1322) is configured with a sorting strategy, wherein the sorting strategy comprises: recommending solutions for the unrecognized custom information in descending order according to the correlation value, wherein the correlation value is calculated according to the correlation value algorithm, and sending the sorting solution to a human for verification, and if the solution verified by the human does not correspond to any of the top three solutions, adjusting the weight in the correlation value algorithm; The correlation value algorithm is configured as: Pxg=a1×Yyb+a2×Tpb+a3×Spb, where Pxg is the correlation value, Yyb is the semantic matching degree of text information, Tpb is the picture matching degree of picture information, Spb is the video matching degree of video information, a1 is the first adjustment weight, a2 is the second adjustment weight, a3 is the third adjustment weight, and a1, a2 and a3 are respectively greater than 0.

2. The community dispute rapid mediation system based on self-learning according to claim 1 is characterized in that: The input module (12) further comprises a custom input unit (122), wherein the custom input unit (122) is used for the user (2) to custom input the type of dispute information; The custom input unit (122) is configured with a custom input strategy, the custom input strategy comprising: dividing the custom input information of the user (2) into a plurality of custom input types, the custom input types comprising a text input type, a picture input type, a video input type and a voice input type; The text input type is configured with a text input box, and the text input box is used for the user (2) to input text information of the dispute information; The picture input type is configured with a picture upload box, and the picture upload box is used for the user (2) to upload picture information of the dispute information; The video input type is configured with a video upload box, and the video upload box is used for the user (2) to upload video information of the dispute information; The voice input type is configured with a voice input box, and the voice input box is used for the user (2) to upload voice information of dispute information.

3. The community dispute rapid mediation system based on self-learning according to claim 2 is characterized in that: The matching unit (131) includes an identification subunit (1311), and the identification subunit (1311) is configured with an identification strategy; the identification strategy includes: identifying the label corresponding to the information selected by the user (2) according to the label of the dispute information; The customized dispute information input by the user (2) is sequentially recognized; semantic extraction is performed on the text information of the text input type, and a piece of dispute information after a label is matched according to the semantics, and the unrecognized text information is marked as unrecognized text information; Identify the image information of the image input type, identify the occurrence scene and the status of the characters in the image information, match a labeled dispute information according to the occurrence scene and the status of the characters, and mark the unrecognized image information as unrecognized image information; Identify the video information of the video input type, identify the occurrence scene and character status of the video information, match a labeled dispute information according to the occurrence scene and character status, and mark the unidentified video information as unidentified video information; Recognize the voice information of voice input type, convert the voice information into text information, extract the semantics of the text information, match a labeled dispute information according to the semantics, and mark the unrecognized voice information as unrecognized voice information.

4. The community dispute rapid mediation system based on self-learning according to claim 3 is characterized in that: The matching unit (131) further comprises a matching subunit (1312), wherein the matching subunit (1312) is configured with a matching strategy, wherein the matching strategy comprises: selecting a solution corresponding to the number from the database module (14) according to the number of the dispute information after the numbering, and sending the solution to the user (2); Manual service is provided for information matching of unrecognized custom input.

Images