Contradictory dispute multivariate solution cross-department collaboration platform and method

Abstract

The invention relates to the technical field of information processing, and discloses a contradictory dispute multivariate solution cross-department collaboration platform and method.The platform comprises a data layer, a processing layer and a collaboration layer which are connected in sequence; the data layer comprises a multi-source data acquisition and fusion module; the processing layer comprises a feature extraction and intelligent identification module; the collaboration layer comprises a cross-department dynamic allocation module; the multi-source data acquisition and fusion module fuses multi-source heterogeneous contradictory dispute data by using an adaptive fusion algorithm; a feature extraction and intelligent identification module performs feature extraction and contradictory dispute type identification on the contradictory dispute description text to obtain a contradictory dispute type; and the cross-department dynamic allocation module constructs a conflict resolution knowledge graph, determines an optimal conflict dispute acceptance department corresponding to the conflict dispute type by using the conflict resolution knowledge graph, and allocates conflict dispute cases to the optimal conflict dispute acceptance department. According to the invention, the processing efficiency and accuracy of contradictory disputes are obviously improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of information processing, and in particular relates to a multi-resolution conflict and dispute resolution cross-departmental collaboration platform and method. BACKGROUND

[0002] With the development and progress of society, various conflicts and disputes inevitably arise, especially in densely populated cities, where various disputes often occur. In order to better resolve these conflicts and disputes, the construction of a conflict and dispute resolution platform is a necessary means.

[0003] In the conflict and dispute resolution process, cases need to be manually classified into mediation, arbitration or litigation departments. This process is time-consuming, error-prone and inefficient. SUMMARY

[0004] Therefore, the present application provides a multi-resolution conflict and dispute resolution cross-departmental collaboration platform and method to solve the problem of time-consuming, error-prone and inefficient conflict and dispute resolution process.

[0005] In a first aspect, the present application provides a multi-resolution conflict and dispute resolution cross-departmental collaboration platform, which comprises a data layer, a processing layer and a collaboration layer connected in sequence; the data layer comprises a multi-source data acquisition and fusion module; the processing layer comprises a feature extraction and intelligent recognition module; the collaboration layer comprises a cross-departmental dynamic allocation module; The multi-source data acquisition and fusion module is used to acquire multi-source heterogeneous conflict and dispute data, and to fuse the multi-source heterogeneous conflict and dispute data using an adaptive fusion algorithm to obtain conflict and dispute description text. The feature extraction and intelligent recognition module is used to perform emotion analysis, time-space analysis and environmental feature analysis on the conflict and dispute description text to obtain a multi-dimensional conflict and dispute feature space, and to perform type recognition based on the multi-dimensional conflict and dispute feature space to obtain a conflict and dispute type. The cross-departmental dynamic allocation module is used to construct a conflict resolution knowledge graph, to determine the optimal conflict and dispute resolution department corresponding to the conflict and dispute type using the conflict resolution knowledge graph, and to allocate conflict and dispute cases to the optimal conflict and dispute resolution department.

[0006] The contradiction dispute multi-resolution cross-department collaborative platform provided by the embodiment fuses multi-source heterogeneous contradiction dispute data, performs batch data processing on the contradiction dispute data corresponding to multiple data sources, and then performs feature extraction on contradiction dispute description text and contradiction dispute type identification, thereby realizing accurate identification of the types corresponding to different contradiction dispute data, and finally determining the optimal contradiction dispute handling department corresponding to the contradiction dispute type by using a conflict resolution knowledge graph, performing contradiction dispute case allocation on the optimal contradiction dispute handling department, constructing a comprehensive contradiction dispute handling process, realizing intelligent handling of contradiction disputes, and significantly improving the processing efficiency and accuracy.

[0007] In an optional implementation, the multi-source data acquisition and fusion module is specifically configured to acquire multi-source heterogeneous contradiction dispute data, perform data preprocessing on the multi-source heterogeneous contradiction dispute data, perform feature extraction on the contradiction dispute data after the data preprocessing, obtain contradiction dispute features, dynamically adjust the weights corresponding to the contradiction dispute data after the data preprocessing, obtain feature weights, and perform weighted fusion on the contradiction dispute features based on the feature weights to obtain contradiction dispute description text.

[0008] The contradiction dispute multi-resolution cross-department collaborative platform provided by the embodiment fuses multi-source heterogeneous contradiction dispute data, performs batch data processing on the contradiction dispute data corresponding to multiple data sources, and then performs feature extraction on contradiction dispute description text and contradiction dispute type identification, thereby realizing accurate identification of the types corresponding to different contradiction dispute data, and finally determining the optimal contradiction dispute handling department corresponding to the contradiction dispute type by using a conflict resolution knowledge graph, performing contradiction dispute case allocation on the optimal contradiction dispute handling department, constructing a comprehensive contradiction dispute handling process, realizing intelligent handling of contradiction disputes, and significantly improving the processing efficiency and accuracy.

[0009] In an optional implementation, the feature extraction and intelligent identification module comprises: The sentiment inclination strengthening analysis unit is configured to perform emotion recognition and voice intonation recognition on the contradiction dispute description text, obtain emotion quantization levels and voice intonation features, associate the emotion quantization levels and the voice intonation features, and obtain emotion features. The text feature enhancement extraction unit is configured to extract contradiction dispute space-time features and obtain social environment features based on the contradiction dispute description text, fuse the emotion features, the contradiction dispute space-time features, and the social environment features, and obtain a multi-dimensional contradiction dispute feature space. The domain adaptive model training unit is configured to obtain historical contradiction dispute case data, perform feature extraction on the historical contradiction dispute case data, determine contradiction dispute general features and contradiction dispute domain features, and perform optimization training on a neural network model by using the historical contradiction dispute case data, the contradiction dispute general features, and the contradiction dispute domain features to obtain a domain adaptive model. The contradiction dispute type identification unit is configured to input the multi-dimensional contradiction dispute feature space into the domain adaptive model to obtain a contradiction dispute type.

[0010] The contradiction dispute multi-resolution cross-department collaborative platform provided in the embodiment combines emotion analysis with space-time characteristics and social environment characteristics, constructs a multi-dimensional contradiction dispute characteristic space, and provides a more comprehensive basis for intelligent allocation of contradiction dispute cases; and by extracting contradiction dispute general characteristics and contradiction dispute field characteristics, the neural network model is optimized and trained to obtain a field adaptive model, which automatically adjusts the focus of characteristics for different types of contradiction disputes, significantly improves the generalization ability and precision of the field adaptive model, and realizes accurate identification of the type of contradiction dispute by using the field adaptive model.

[0011] In an optional implementation, the cross-department dynamic allocation module comprises: The dynamic capability evaluation unit is configured to obtain contradiction dispute handling indexes of a plurality of contradiction dispute handling departments, and construct a dynamic department portrait based on the contradiction dispute handling indexes of the plurality of contradiction dispute handling departments. The conflict resolution unit is configured to construct a regulation knowledge graph, determine a regulation path corresponding to the type of contradiction dispute by using the regulation knowledge graph, calculate semantic similarity based on the contradiction dispute text corresponding to the type of contradiction dispute and the regulation clauses in the regulation path, and obtain conflict resolution rules, and determine the optimal contradiction dispute handling department based on the semantic similarity, the conflict resolution rules, and the dynamic department portrait. The collaborative processing flow optimization unit is configured to establish a cross-department collaborative processing standardized process architecture, and allocate the contradiction dispute case by using the cross-department collaborative processing standardized process architecture based on the optimal contradiction dispute handling department.

[0012] The contradiction dispute multi-resolution cross-department collaborative platform provided in the embodiment determines the optimal contradiction dispute handling department based on semantic similarity, conflict resolution rules, and a dynamic department portrait, effectively solving the cross-department jurisdiction dispute.

[0013] In an optional implementation, the system further comprises a security layer connected with the data layer, the processing layer, and the collaboration layer; the security layer comprises a data encryption and access control module; the data encryption and access control module comprises a federated learning data collaboration framework construction unit and a multi-level encryption unit. The federated learning data collaboration framework construction unit is configured to obtain contradiction dispute handling departments participating in contradiction dispute handling, take the contradiction dispute handling departments participating in contradiction dispute handling as clients, and construct a federated learning data collaboration framework based on a central server and a plurality of clients; the central server is connected with the plurality of clients. The multi-level encryption unit is used to obtain local contradiction dispute data of the client, train and update a local contradiction dispute classification model based on the local contradiction dispute data of the client, obtain updated model parameters, encrypt the updated model parameters, obtain encryption parameters corresponding to the client, aggregate the encryption parameters through the central server, obtain global contradiction dispute classification model parameters, update the global contradiction dispute classification model by using the global contradiction dispute classification model parameters, encrypt the updated global contradiction dispute classification model, and send the encrypted global contradiction dispute classification model to the client for contradiction dispute classification processing.

[0014] The contradiction dispute multi-resolution cross-department collaborative platform provided in the embodiment constructs a federated learning data collaboration framework, shares the encrypted model parameters by using the federated learning data collaboration framework, realizes data sharing of multiple contradiction dispute handling departments under the premise of guaranteeing data security, and expands the system data dimension and analysis depth.

[0015] In an optional implementation, the contradiction dispute multi-resolution cross-department collaborative platform further includes a display layer connected with the data layer, the processing layer and the collaboration layer respectively; and the display layer includes a user interaction and visualization module. The user interaction and visualization module is configured to collect and store contradiction dispute handling data, and perform multi-terminal access control and customized configuration on the contradiction dispute handling data.

[0016] The contradiction dispute multi-resolution cross-department collaborative platform provided in the embodiment realizes visualized tracing of the whole process of contradiction dispute handling by performing multi-terminal access control and customized configuration on the contradiction dispute handling data, and improves the transparency and public credibility of the system.

[0017] In a second aspect, the contradiction dispute multi-resolution cross-department collaborative method is applied to the contradiction dispute multi-resolution cross-department collaborative platform in the first aspect or any possible implementation thereof, and the method includes the following steps. The multi-source data collection and fusion module collects multi-source heterogeneous contradiction dispute data, and fuses the multi-source heterogeneous contradiction dispute data by using an adaptive fusion algorithm to obtain contradiction dispute description text. The feature extraction and intelligent identification module performs emotion analysis, space-time analysis and environmental feature analysis on the contradiction dispute description text to obtain a multi-dimensional contradiction dispute feature space, and performs type identification based on the multi-dimensional contradiction dispute feature space to obtain a contradiction dispute type. The cross-department dynamic allocation module constructs a conflict resolution knowledge graph, determines an optimal contradiction dispute handling department corresponding to the contradiction dispute type by using the conflict resolution knowledge graph, and allocates contradiction dispute cases to the optimal contradiction dispute handling department.

[0018] In a third aspect, the present application provides a computer device, comprising: a memory and a processor, which are connected in communication with each other, and the memory stores computer instructions; the processor executes the computer instructions to perform the contradiction dispute multi- resolution cross-department collaboration method of the second aspect.

[0019] In a fourth aspect, the present application provides a computer readable storage medium, which stores computer instructions for making a computer execute the contradiction dispute multi- resolution cross-department collaboration method of the second aspect.

[0020] In a fifth aspect, the present application provides a computer program product, which comprises computer instructions for making a computer execute the contradiction dispute multi- resolution cross-department collaboration method of the second aspect. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the specific embodiments or the prior art in the present application, the drawings needed to be used in the specific embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0022] Figure 1 is a structural block diagram of a contradiction dispute multi-resolution cross-department collaboration platform according to an embodiment of the present application; Figure 2 is a structural block diagram of a feature extraction and intelligent identification module according to an embodiment of the present application; Figure 3 is a structural block diagram of a cross-department dynamic allocation module according to an embodiment of the present application; Figure 4 is a structural block diagram of a data encryption and access control module according to an embodiment of the present application; Figure 5 is a flowchart of a contradiction dispute multi-resolution cross-department collaboration method according to an embodiment of the present application; Figure 6 is a hardware structure schematic diagram of a computer device of an embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0024] It can be understood that, before using the technical solutions disclosed in the embodiments of the present application, the type, use range, use scenario and the like of the personal information involved in the present application should be informed to the user and the authorization of the user should be obtained through appropriate means according to relevant laws and regulations.

[0025] For example, in response to receiving the active request of the user, prompt information is sent to the user to explicitly prompt the user that the operation requested to be performed will require obtaining and using the personal information of the user. Thus, the user can voluntarily choose whether to provide the personal information to the software or hardware such as the electronic device, application program, server or storage medium performing the operation of the technical solution of the present application according to the prompt information.

[0026] As an optional but non-limiting implementation manner, in response to receiving the active request of the user, the manner of sending the prompt information to the user may, for example, be a pop-up window manner, and the prompt information may be presented in the form of text in the pop-up window. In addition, the pop-up window may also carry selection controls for the user to select "agree" or "disagree" to provide the personal information to the electronic device.

[0027] It can be understood that the above notification and user authorization process is only illustrative, and does not limit the implementation manner of the present application, and other manners meeting the relevant laws and regulations can also be applied to the implementation manner of the present application.

[0028] It can be understood that the data involved in the present technical solution (including but not limited to the data itself, the acquisition or use of the data) should comply with the requirements of the relevant laws and regulations and relevant provisions.

[0029] The relevant contradiction dispute resolution platform aims to solve social contradictions and disputes through various means and channels, and maintain social harmony and stability. However, in the work practice, there are also some problems and deficiencies: 1. Lack of informatization and intelligentization: With the development of information technology, the handling of contradictions and disputes should also be assisted by informatization and intelligentization means to improve the processing efficiency and quality.

[0030] 2. Lack of professionalism and skill: The staff lack professional knowledge and skills, and cannot effectively handle contradictions and disputes. Some places and units also lack professional contradiction and dispute handling institutions and personnel, resulting in low efficiency of contradiction and dispute handling.

[0031] Three, lack of scientificity and systematicness: the platform only solves the current dispute, lacks long-term planning and systematic solution. This is easy to lead to the repeated occurrence of contradictions and disputes, and even exacerbate the intensification of contradictions.

[0032] To solve the above problems, the embodiment of the application provides a multi-dispute resolution cross-department collaboration platform, which automatically classifies and distributes according to case characteristics, and has real-time monitoring and early warning functions, significantly improving processing efficiency and accuracy; wherein the multi-dispute resolution cross-department collaboration platform can also be applied to the following scenarios: 1) Group event early warning: through spatiotemporal feature analysis and emotional heat map, potential group disputes are early warned and preventive intervention is made; for example, it is monitored that the labor dispute emotion value around a certain enterprise continues to rise in the short term, and a possible group event is early warned.

[0033] 2) Cross-regional dispute processing: use knowledge graph to associate cross-regional party relationships, and intelligently recommend the best cross-regional collaborative processing scheme; for example, recommend a joint mediation scheme for a contract dispute involving suppliers and purchasers in multiple places.

[0034] 3) Intelligent law and regulation push: push relevant regulations and typical cases in real time according to dispute characteristics to assist mediation and arbitration; for example, in a lease contract dispute, push local housing lease regulations and similar case judgments.

[0035] 4) Mobile terminal convenient processing: develop a mobile terminal application to support on-site evidence collection and instant messaging mediation functions; for example, take a photo of a traffic accident dispute on site and upload it to instantly generate a liability determination suggestion.

[0036] 5) Third-party agency collaboration: access notaries, appraisers and other third-party agencies to build a one-stop dispute resolution ecosystem; for example, in a copyright infringement dispute, automatically connect to a judicial appraisal agency to obtain professional appraisal opinions.

[0037] The embodiment provides a multi-dispute resolution cross-department collaboration platform, as shown in Figure 1 Fig. 1, which comprises a data layer 101, a processing layer 102 and a collaboration layer 103 connected in sequence; the data layer comprises a multi-source data acquisition and fusion module 1011; the processing layer comprises a feature extraction and intelligent identification module 1021; the collaboration layer comprises a cross-department dynamic allocation module 1031; The multi-source data acquisition and fusion module 1011 is used to acquire multi-source heterogeneous dispute data, fuse the multi-source heterogeneous dispute data by using an adaptive fusion algorithm, and obtain a dispute description text.

[0038] Specifically, the multi-source data acquisition and fusion module 1011 supports real-time data access and batch data processing.

[0039] The feature extraction and intelligent recognition module 1021 is configured to perform emotion analysis, space-time analysis and environmental feature analysis on the dispute description text, to obtain a multi-dimensional dispute feature space, and to perform type recognition based on the multi-dimensional dispute feature space, to obtain a dispute type.

[0040] The cross-department dynamic allocation module 1031 is configured to construct a conflict resolution knowledge graph, to determine an optimal dispute handling department corresponding to the dispute type by using the conflict resolution knowledge graph, and to allocate dispute cases to the optimal dispute handling department.

[0041] The multi-dispute resolution cross-department collaboration platform provided in this embodiment fuses multi-source heterogeneous dispute data, performs batch data processing on the dispute data corresponding to various data sources, and then performs feature extraction and dispute type recognition on the dispute description text, to accurately recognize the types of different dispute data. Finally, the conflict resolution knowledge graph is used to determine an optimal dispute handling department corresponding to the dispute type, and dispute cases are allocated to the optimal dispute handling department, to construct a comprehensive dispute handling process, realize intelligent dispute handling, and significantly improve the processing efficiency and accuracy.

[0042] In some optional embodiments, the multi-source data acquisition and fusion module 1011 is specifically configured to acquire multi-source heterogeneous dispute data, perform data preprocessing on the multi-source heterogeneous dispute data, perform feature extraction on the preprocessed dispute data, obtain dispute features, dynamically adjust the weights of the preprocessed dispute data, obtain feature weights, perform weighted fusion on the dispute features based on the feature weights, and obtain dispute description text.

[0043] Specifically, the dispute data is fused with external data sources (data of different types, different fields, different formats, etc.) such as industry databases, an adaptive fusion algorithm is designed to dynamically adjust the weights, and the overall situation of the dispute is more accurately reflected; for example, the transaction flow data in financial disputes is combined with the credit score of the parties, to improve the classification accuracy of financial disputes.

[0044] Further, data is collected from different sources such as court litigation systems, labor arbitration agencies, people's mediation committees and industry databases, and the data types cover structured data (such as case handling numbers, basic information of parties, etc.), semi-structured data (dispute description text, chat records, etc.) and unstructured data (on-site photos, audio and video recordings, etc.); taking financial disputes as an example, transaction flow data (including transaction time, amount, transaction parties, etc.) is obtained from a bank transaction system, and credit score data of parties is obtained from a credit evaluation agency.

[0045] Further, duplicate data is removed, such as information that is repeatedly entered in different department systems for the same case. For missing values, mean filling (for numerical data) or mode filling (for categorical data) is used for supplementation; for abnormal data, correction is made, such as negative transaction amount or far beyond the reasonable range, according to business rules for correction or marking as suspicious data for subsequent verification; the field format of data from different sources is uniformly standardized; for example, the date format is unified as "YYYY-MM-DD", and the amount unit is unified as "yuan".

[0046] Further, the steps of initial weight setting and dynamic adjustment include: determining the category of data sources and their importance: based on historical data statistics and expert experience, the importance of different data sources in dispute classification is preliminarily evaluated, for example, in financial disputes, transaction flow data and party credit score data are key factors; initial weight assignment: based on the importance evaluation results, initial weights are assigned to each type of data source, such as transaction flow data initial weight set to 0.4, party credit score initial weight set to 0.3, dispute description text weight set to 0.2, and other data source weights set in turn; building a weight adjustment model: using regression algorithms or neural networks in machine learning, a weight adjustment model is established, taking the processing results (such as classification accuracy) of historical dispute cases as the supervision signal, and the features input into the model include the initial weights of each data source, data quality indicators (such as data integrity, accuracy, etc.) and dispute type related features; dynamic weight calculation: in actual application, for each new dispute case, the initial weights of each data source and related feature data are input into the weight adjustment model, and the model dynamically calculates the optimal weight of each data source in the current case according to the rules learned from historical data, for example, when dealing with a dispute involving complex financial derivatives, the model may increase the weight of related industry database data, because such data is more critical to accurate classification in this specific scenario.

[0047] Further, feature extraction is performed on each data source. For transaction flow data, transaction frequency, single transaction amount, transaction counterpart nature and other features are extracted; for party credit score data, credit rating, overdue record times and other features are extracted; for dispute description text, natural language processing techniques are used to extract keywords, text vectors and other features.

[0048] Further, according to the dynamically calculated weights, the features of each data source are weighted and fused; for example, the transaction flow feature vector is multiplied by the corresponding weight, the credit score feature vector is multiplied by the weight, and the text feature vector is multiplied by the weight, then these weighted feature vectors are spliced or fused to calculate a fusion vector that comprehensively reflects the dispute features, i.e. the contradiction dispute description text.

[0049] The contradiction dispute multi-resolution cross-departmental collaborative platform provided by the embodiment utilizes an adaptive fusion algorithm to fuse contradiction dispute data, realizes dynamic weight adjustment of multi-source heterogeneous contradiction dispute data, and improves the accuracy of contradiction dispute feature extraction, especially in complex cases.

[0050] In some optional embodiments, as shown in FIG. 10, the feature extraction and intelligent identification module 1021 includes: Figure 2 The emotion tendency strengthening analysis unit 10211 is configured to perform emotion recognition and voice intonation recognition on the contradiction dispute description text, obtain an emotion quantization level and a voice intonation feature, and associate the emotion quantization level and the voice intonation feature to obtain an emotion feature.

[0051] Specifically, the emotion analysis technology in natural language processing is used to perform emotion recognition on the dispute description text: an emotion classification system is established, and the emotion is classified into five basic emotions, such as anger, anxiety, sadness, joy, and fear. For example, in a family dispute text, the anger emotion is recognized by analyzing keywords such as "throwing things" and "shouting quarrels", and the anxiety emotion is recognized by analyzing keywords such as "insomnia" and "low mood".

[0052] Further, each emotion type is quantized and graded into 1-5 levels, that is, the frequency, intensity, and sentence structure of the emotion keywords in the text are used to make a grading judgment. Taking the anger emotion as an example, level 1 represents slight dissatisfaction, and level 5 represents extreme anger.

[0053] Further, the voice input of the parties in the contradiction dispute description text is preprocessed, including voice framing, endpoint detection, and other operations, and then the voice intonation features such as the fundamental frequency (reflecting the pitch change) and the sound intensity (reflecting the volume) are extracted. For example, the fundamental frequency curve of the voice signal is extracted by a voice processing tool to analyze its trend; the mean, variance, and other statistical indicators of the sound intensity are calculated.

[0054] Further, a correlation model between the voice intonation feature and the emotion is established: by analyzing a large amount of voice data with emotion labels, a machine learning model (such as a decision tree, a support vector machine, etc.) is trained to learn the distribution of voice intonation features under different emotional states. When new voice data is input, the model is used to determine the emotional state and intensity of the parties according to the voice intonation features.

[0055] The text feature enhancement extraction unit 10212 is configured to extract contradiction dispute space-time features based on the contradiction dispute description text, obtain social environment features, fuse the emotion features, the contradiction dispute space-time features, and the social environment features, and obtain a multi-dimensional contradiction dispute feature space.

[0056] ​Specifically, the address information of the contradiction dispute case in the contradiction dispute description text is converted into three-dimensional geographic coordinates (longitude, latitude, and altitude), and the geographic information system (GIS) technology is used to process and analyze the geographic coordinates in combination with map data to obtain spatial features. For example, the environmental features around the case occurrence place are analyzed, such as whether it is located in a commercial area, a residential area, or near a school, etc. The above information may be related to the type and nature of the dispute.

[0057] Further, the case occurrence time is finely divided to the minute, and time-related features such as time period (day, night), day of the week, holiday, etc. are extracted. For example, the frequency of occurrence of disputes in different time periods is statistically analyzed to find that certain types of disputes (such as entertainment venue consumption disputes) have a higher frequency of occurrence at night.

[0058] Further, the extracted spatial features, time features, and previously obtained emotional features are fused. A space-time-emotion feature matrix can be constructed, with the rows representing different cases and the columns representing the fused feature dimensions. For example, a case in a commercial area (spatial feature), at 9 pm (time feature), and with a party's emotional anger level of 4 (emotional feature) are combined into a feature vector as the comprehensive space-time-emotion feature representation of the case.

[0059] Further, convolutional neural networks are used to extract space-time correlation features, accurately locate the space-time pattern of the dispute, and introduce macroeconomic indicators, policy changes, and other social environmental features to obtain a multi-dimensional contradiction dispute feature space.

[0060] The domain adaptation model training unit 10213 is configured to obtain historical contradiction dispute case data, extract features from the historical contradiction dispute case data, determine general contradiction dispute features and domain-specific contradiction dispute features, optimize and train the neural network model using the historical contradiction dispute case data, the general contradiction dispute features, and the domain-specific contradiction dispute features, and obtain a domain adaptation model.

[0061] Specifically, a large number of historical contradiction dispute case data of different types are collected, which are labeled according to the type of dispute, such as contract disputes, labor disputes, and tort disputes. Features are extracted from each case data, including text features, party features, and case fact features. For example, in contract dispute data, features such as contract clause content, performance, and breach of contract are extracted. In labor dispute data, features such as wage payment, working conditions, and working hours system are extracted.

[0062] Further, a neural network model suitable for text classification and feature learning is selected, such as a convolutional neural network (CNN) or a long short-term memory network (LSTM); or, an LSTM is used to process text sequence features, while a CNN is used to extract case element spatial features, feature-level fusion is achieved through an attention mechanism to improve dispute type recognition accuracy; the labeled historical contradictory dispute case data is input into the model for training, so that the model learns the general feature representation of different dispute type data. During the training process, the performance of the model is optimized by adjusting the hyperparameters of the model (such as learning rate, batch size, regularization parameter, etc.), to improve the classification accuracy.

[0063] Further, for specific dispute fields, such as labor dispute fields, in-depth analysis of the unique features of the field is performed, i.e., through expert knowledge and text mining technology, labor law-related provisions, and special terms of labor relations (such as “overtime pay” and “social insurance”) are identified as field features; in the field of contract disputes, key provisions in relevant contract law, contract types (such as sales contracts and lease contracts), and specific expressions are mined as field features.

[0064] Further, a domain adaptation module is introduced into the above neural network model, which includes a feature weight adjustment subnetwork. When processing dispute data in different fields, the general features of contradictory disputes and the field-specific features of contradictory disputes are input into the subnetwork to calculate the attention weight of the field features, and then the trained neural network model is used as a domain adaptation model; for example, when processing labor dispute data, the subnetwork will increase the attention weight of the labor law-related features (such as wage payment provisions); when processing contract dispute data, the weight of the contract law-related features (such as breach of contract provisions) is increased.

[0065] Further, an online learning mechanism is established. When new dispute case data is input, the domain adaptation model is updated in real time, and the parameters of the model are fine-tuned according to the field features and processing results of the new data; for example, when a new labor dispute case is correctly classified, the domain adaptation model feeds back the feature information of the case to the domain adaptation module, strengthening the learning and weight adjustment of the labor dispute field features.

[0066] Further, the adaptability of the domain adaptation model in different fields is evaluated periodically, and the performance of the model in each field is analyzed by calculating the classification accuracy, recall rate, F1 value (a machine learning evaluation index used to comprehensively evaluate the performance of a classification model), and other indicators. For fields with poor performance, the model structure or training strategy is adjusted accordingly, such as increasing the training data in that field, optimizing the feature extraction method, etc.

[0067] The conflict and dispute type identification unit 10214 is used to input the multi-dimensional conflict and dispute feature space into the domain adaptive model to obtain the conflict and dispute type.

[0068] Specifically, the core algorithm code for identifying conflict types is as follows: Function dispute_type_identification(text_data,spatiotemporal_data,social_data): Text feature extraction text_features = lstm_layer(text_data) Spatiotemporal feature extraction spatiotemporal_features = cnn_layer(spatiotemporal_data) Social environment feature extraction social_features = dense_layer(social_data) Feature fusion fused_features=attention_mechanism(text_features,spatiotemporal_features,social_features) Dispute type prediction type_prediction = dense_layer(fused_features) return type_prediction This embodiment provides a cross-departmental collaborative platform for diversified dispute resolution, which combines sentiment analysis with spatiotemporal and social environmental features to construct a multi-dimensional feature space for disputes, providing a more comprehensive basis for the intelligent allocation of dispute cases. Furthermore, by extracting general and domain-specific features of disputes, the neural network model is optimized and trained to obtain a domain-adaptive model. This model automatically adjusts the focus of feature selection for different types of disputes, significantly improving the generalization ability and accuracy of the domain-adaptive model. The accurate identification of dispute types is achieved using the domain-adaptive model.

[0069] In some alternative implementations, such as Figure 3 As shown, the department dynamic allocation module 1031 includes: The dynamic capability evaluation unit 10311 is configured to acquire the contradiction dispute handling indexes of a plurality of contradiction dispute accepting departments, and construct a dynamic department portrait based on the contradiction dispute handling indexes of the plurality of contradiction dispute accepting departments.

[0070] Specifically, the processing efficiency (case / hour), resource occupancy rate and other indexes of each contradiction dispute accepting department are acquired in real time, and a dynamic department capability portrait is constructed to realize accurate case allocation; for example, the average time of an arbitration court to hear a complex case is predicted based on historical data, and new cases are reasonably allocated.

[0071] The conflict resolution unit 10312 is configured to construct a regulation knowledge graph, determine a regulation path corresponding to the type of the contradiction dispute by using the regulation knowledge graph, calculate a semantic similarity between the contradiction dispute text corresponding to the type of the contradiction dispute and the regulation clauses in the regulation path, and acquire a conflict resolution rule, and determine an optimal contradiction dispute accepting department based on the semantic similarity, the conflict resolution rule and the dynamic department portrait.

[0072] Specifically, when multiple departments simultaneously claim jurisdiction, a conflict resolution procedure based on semantic similarity and jurisdiction regulation knowledge graph is started to automatically determine the optimal contradiction dispute accepting department and avoid departmental shirking; for example, when a labor dispute involves tort, the key feature weight is distributed to the arbitration or mediation department.

[0073] Further, the regulation texts related to dispute handling, such as national laws and regulations, local regulations and departmental rules, are acquired, and the regulation texts are sorted and labeled by chapters and clauses; the named entity recognition and relationship extraction methods in the natural language processing technology are used to extract legal entities (such as “labor contract”, “employer”, “laborer”, “breach of contract liability” and “tort liability”) and the relationships between the entities (such as the “signing” relationship between “labor contract” and “employer”, and the “basis” relationship between “breach of contract liability” and “contract clauses”) from the regulation texts; a pre-trained language model is used for identification and classification of entities and relationships to improve the accuracy and efficiency of extraction; the pre-trained language model can be a BERT (Bidirectional Encoder Representations from Transformers, pre-trained language model based on Transformer architecture) model.

[0074] Further, the extracted entities and relations are constructed into a regulation knowledge graph, stored in the form of a graph structure, where the nodes represent legal entities and the edges represent the relationships between entities. A graph database (such as Neo4j) is used for storage and management, facilitating subsequent queries and reasoning. For example, in the knowledge graph, a "labor dispute" node and a "labor arbitration" node establish a "jurisdiction" relationship edge, and a "labor arbitration" node and a "people's court" node establish a "litigation connection" relationship edge.

[0075] Further, for the case description text and the regulation clause text declared by each dispute acceptance department, preprocessing operations are performed, including word segmentation, stop word removal, and stem extraction. Then, word embedding techniques (such as Word2Vec, GloVe, etc.) are used to convert the text into vector representations. For example, the case description "laborer and employer have a dispute over overtime pay" is converted into a word vector sequence, and the relevant regulation clause "the employer shall pay the laborer timely and sufficient labor remuneration according to the labor contract agreement and national regulations" is also converted into a word vector sequence.

[0076] Further, a suitable semantic similarity calculation method is selected, such as cosine similarity, Jaccard similarity (a measure of the similarity between two sets), or a deep learning-based similarity calculation model (such as Siamese network). The cosine similarity is used to calculate the similarity between the case text and the above regulation clause text. The formula for calculating the cosine similarity is:

[0077] where, and are two text vectors (i.e., case text and regulation clause text), represents the dot product of the vectors, and represent the length of the vectors and , respectively.

[0078] Further, according to historical data and expert experience, a semantic similarity threshold is set, such as 0.6. When the similarity between the case description text and a certain regulation clause text is higher than the threshold, it is considered that the regulation clause is related to the case and may be used as a basis for determining jurisdiction. For example, the similarity between the case description and multiple regulation clauses is calculated, and it is found that the similarity with the labor remuneration payment clause is 0.7, which is higher than the threshold. Therefore, this regulation clause is included in the consideration range.

[0079] Further, according to laws and regulations and judicial practice experience, a set of conflict resolution rules are formulated, which include but are not limited to the principles of "special law prior to general law", "later law prior to earlier law", "superior law prior to inferior law", etc.; for example, when a labor dispute case involves both labor law and local labor regulations, according to the principle of "superior law prior to inferior law", the relevant provisions in the labor law are preferred.

[0080] Further, reasoning is performed using the regulation knowledge graph, the entities and relationship paths related to the contradiction dispute case are found in the regulation knowledge graph, the optimal contradiction dispute acceptance department is determined in combination with the semantic similarity calculation result and the conflict resolution rule, and meanwhile, the optimal contradiction dispute acceptance department meets the above dynamic department portrait; for example, for a case involving both labor dispute and tort liability, it is found through regulation knowledge graph reasoning that the labor dispute part has stronger relevance with the relevant labor arbitration clauses, and the tort liability part has stronger relevance with the relevant tort liability clauses, according to the conflict resolution rule, it is determined that the labor arbitration department has priority in acceptance, and if the arbitration result involves the part of tort compensation that cannot be solved, it is guided to the court litigation department, wherein the labor arbitration department and the court litigation department meet the corresponding dynamic department portrait.

[0081] The collaborative processing flow optimization unit 10313 is configured to establish a cross-department collaborative processing standardized process architecture, and distribute the contradiction dispute case based on the optimal contradiction dispute acceptance department using the cross-department collaborative processing standardized process architecture.

[0082] Specifically, the cross-department collaborative processing standardized process is designed to clearly define the responsibilities and interaction nodes of each department, and to improve the overall processing efficiency; for example, mediation failure is automatically transferred to arbitration, and arbitration results are not satisfied and are guided to litigation, and the whole process is seamlessly connected.

[0083] The contradiction dispute multi-resolution cross-department collaborative platform provided in this embodiment determines the optimal contradiction dispute acceptance department based on semantic similarity, conflict resolution rules and dynamic department portraits, and effectively solves the cross-department jurisdiction dispute.

[0084] In some optional embodiments, further comprising: a security layer 104 connected with the data layer 101, the processing layer 102 and the collaboration layer 103 respectively; the security layer 104 includes a data encryption and access control module 1041; as shown in Figure 4 The data encryption and access control module 1041 includes a federated learning data collaboration framework construction unit 10411 and a multi-level encryption unit 10412. The federal learning data collaboration framework construction unit 10411 is configured to acquire a dispute handling department participating in dispute handling, take the dispute handling department participating in dispute handling as a client, construct a federal learning data collaboration framework based on a central server and multiple clients, and connect the central server with the multiple clients.

[0085] Specifically, multiple departments participating in data collaboration, such as a court, a public security bureau, a judicial bureau, and a labor arbitration committee, are determined, a network architecture of federal learning is designed, and the network architecture includes a central server and multiple clients (each dispute handling department as a client). The central server is responsible for coordinating communication between the clients, aggregating model parameters, and the like. The clients are responsible for local data processing and model training.

[0086] Further, a secure communication protocol is established to ensure the confidentiality and integrity of data during transmission. Encryption technology is used to encrypt transmitted data and model parameters. The encryption technology can use SSL (Security Socket Layer, a security protocol technology for encrypted transmission of network communication data based on public key cryptography) / TLS (Transport Layer Security, a transmission layer security protocol) encryption. For example, when transmitting model parameters between the client and the central server, an asymmetric encryption algorithm is used for encryption. Only the central server has the decryption private key, ensuring that the parameters cannot be stolen or tampered with during transmission.

[0087] The multi-level encryption unit 10412 is configured to acquire local dispute data of the client, train and update a local dispute classification model based on the local dispute data of the client, obtain updated model parameters, encrypt the updated model parameters to obtain encryption parameters corresponding to the client, aggregate the encryption parameters through the central server to obtain global dispute classification model parameters, update a global dispute classification model using the global dispute classification model parameters, encrypt the updated global dispute classification model, and send the encrypted global dispute classification model to the client for dispute classification processing.

[0088] Specifically, each dispute handling department locally preprocesses its own data, including data cleaning, feature selection, and the like. The data is converted into a unified format and dimension according to the requirements of the federal learning task. For example, the court performs word segmentation processing on the text description in the litigation case data to extract keyword features. The public security bureau performs structured processing on the information of parties involved in the public security dispute data and the course of events.

[0089] Further, at the client side, the initial model is trained using the preprocessed data, a machine learning algorithm suitable for the characteristics of the local data (such as logistic regression, random forest, etc.) is used to update the parameters of the local dispute classification model; for example, the court client uses local litigation data to train a dispute classification model, and by adjusting the weight and bias parameters of the model, the model achieves good classification results on local data.

[0090] Further, the model parameters updated locally by the client are encrypted, and homomorphic encryption technology is used to enable the central server to perform aggregation operations without decrypting the parameters. The encrypted parameters are uploaded to the central server through a secure communication protocol; for example, the Paillier homomorphic encryption algorithm is used to encrypt the model parameters, and the encrypted parameters are sent to the central server in ciphertext form. The Paillier homomorphic encryption algorithm is a probabilistic public key encryption system based on the composite residue class difficulty problem, supporting additive homomorphism and number multiplication homomorphism operations.

[0091] Further, the central server selects a suitable parameter aggregation method according to the federated learning algorithm, such as the Federated Averaging (FedAvg) algorithm, which assigns weights to each client according to the size of the client's data, and calculates the global dispute classification model parameters; for example, if the data volume of the court client accounts for 30% of the total data volume of all participating departments, the model parameter weight of the court client is 0.3 during aggregation, and other clients are calculated accordingly.

[0092] Further, the central server updates the global dispute classification model using the aggregated global dispute classification model parameters, and distributes the updated global dispute classification model to each client in an encrypted manner. Each client uses the new global dispute classification model as the initial model for the next round of training; for example, after the central server completes parameter aggregation, it obtains a new global dispute classification model, encrypts it and sends it to the court, public security bureau and other clients. Each client decrypts and uses the new global dispute classification model for further training on local data.

[0093] Further, the performance of the global dispute classification model is evaluated periodically, and the global dispute classification model is tested on local data at each client to calculate classification accuracy, recall rate and other indicators. Based on the evaluation results, the hyperparameters of the federated learning (such as learning rate, aggregation period, etc.) are adjusted to optimize the model performance until the desired target is achieved.

[0094] Further, the above local dispute classification model and global dispute classification model can use a domain adaptive model.

[0095] Further, when multiple departments share data, federated learning is used to ensure that data does not leave the domain, and only model parameters are encrypted and shared, which meets the requirements of data localization regulations; for example, public security and judicial departments cooperate to handle public security disputes, and data encryption and sharing improve processing efficiency while protecting privacy.

[0096] Further, the key node data of dispute handling is stored in a blockchain to ensure that the data cannot be tampered with, and to provide a trusted evidence chain for judicial traceability; for example, the signing process of the mediation agreement is stored to ensure the effectiveness of subsequent execution.

[0097] Further, sensitive information is anonymized before data analysis and model training to balance data utilization and privacy protection; for example, personal identification is removed, and only dispute features are retained for model training.

[0098] Further, a full-process audit mechanism for data access and operation behavior is established, and a security audit report is generated regularly to timely discover and rectify security vulnerabilities; for example, each data query operation is recorded, and abnormal access behavior is warned in real time.

[0099] The contradiction dispute multi-resolution cross-department collaboration platform provided in this embodiment constructs a federated learning data collaboration framework, uses the federated learning data collaboration framework to share encrypted model parameters, realizes data sharing among multiple contradiction dispute handling departments under the premise of ensuring data security, and expands the system data dimension and analysis depth.

[0100] In some optional embodiments, it further includes a display layer 105 connected with the data layer 101, the processing layer 102 and the collaboration layer 103; the display layer 105 includes a user interaction and visualization module 1051. The user interaction and visualization module 1051 is used to collect and store contradiction dispute handling data, and perform multi-terminal access control and custom configuration on the contradiction dispute handling data.

[0101] Specifically, data collection nodes are set at each link of contradiction dispute handling, including case acceptance, information input, distribution, process record, result feedback, etc., and detailed information of each node is collected by using an automatic tool and manual input in combination; for example, at the case acceptance node, information such as case source (telephone, network, site, etc.), acceptance time, acceptance personnel, etc. is recorded; at the process record node, the conversation content in the mediation process, the court record in the arbitration process, the evidence presentation situation in the litigation process, etc. are recorded.

[0102] Further, a distributed database system is established to store the full-process case data. The database combines relational databases (such as MySQL) and non-relational databases (such as MongoDB) to adapt to the storage needs of structured and unstructured data. For example, case basic information (case number, party information, etc.) is stored in a relational database, and text records, pictures, videos, and other unstructured data in the processing process are stored in a non-relational database, and are associated through the case number.

[0103] Further, permissions are set according to different user roles (case parties, mediators, arbitrators, judges, management personnel, etc.), and each role can see different data and operation functions on the visualization interface. For example, case parties can only view basic information and processing progress related to their own cases; mediators can view and input detailed information during the mediation process; management personnel can view overall statistical information of all cases and system operation status.

[0104] Further, the full process of a case from acceptance to completion is visually presented in the form of a flowchart, and key information such as processing time, processing personnel, and processing results is displayed at each link. For example, different colored nodes in the flowchart represent different processing stages (acceptance, mediation, arbitration, litigation), and clicking on a node allows you to view detailed information for that stage, such as conversation records and mediation plans in the mediation stage.

[0105] Further, the full-process data is summarized and analyzed: a variety of statistical charts (bar charts, line charts, pie charts, etc.) are provided to display case processing efficiency, proportion of different types of disputes, and workload of each department. For example, a bar chart shows the number of different types of conflicts accepted each month, a line chart shows the trend of average case processing time, and a pie chart shows the proportion of the number of cases handled by each department.

[0106] Further, a data correlation index is established to correlate data from each link through case numbers, party information, and other key fields, and an efficient query algorithm is developed to allow users to quickly query full-process information based on different query conditions (such as case number, party name, processing time range, etc.). For example, after the user inputs the case number, the system quickly locates the data records of the case at each link through the correlation index and displays them in an integrated form on the visualization interface.

[0107] Further, a data backup plan is developed, and full-process data is backed up regularly, using multiple copy backup and off-site disaster recovery strategies to ensure data security and reliability; for example, full backup of the database is performed every morning, and backup data is stored in an off-site data center, so that in the event of data loss or damage, the system can be recovered from the backup data in a timely manner to ensure the normal operation of the traceability function.

[0108] Further, all user operations on the visualization interface are recorded, including queries, modifications, deletions, and other operations, an audit mechanism is established, and operation logs are reviewed regularly to ensure data integrity and system security; for example, operation logs record user login times, query contents, operation times, and other detailed information, and auditors review operation logs to identify abnormal operation behaviors (such as unauthorized bulk deletion of data) and handle them in a timely manner.

[0109] The multi-resolution conflict dispute resolution cross-departmental collaboration platform provided in this embodiment realizes visual traceability of the full process of conflict dispute resolution through multi-terminal access control and custom configuration of conflict dispute resolution data, and improves system transparency and credibility.

[0110] The working process of a multi-resolution conflict dispute resolution cross-departmental collaboration platform will be described below through a specific embodiment.

[0111] Embodiment 1: The working process of a multi-resolution conflict dispute resolution cross-departmental collaboration platform includes: I. Special preprocessing of dispute data: 1) Multi-source heterogeneous data fusion: Fusion of dispute data and external data sources (data of different types, different fields, different formats, etc.) such as industry databases, design of adaptive fusion algorithm to dynamically adjust weights, and more accurate reflection of the overall picture of the dispute; for example, combining transaction flow data in financial disputes with party credit scoring to improve the accuracy of financial dispute classification.

[0112] 2) Emotion tendency enhancement analysis: Introduce emotion granularity subdivision (5 levels of anger, anxiety, etc.), combine voice tone analysis to build emotion intensity heat maps, and assist in intelligent allocation; for example, in family disputes, emotion analysis can help identify potential domestic violence risks and intervene in advance.

[0113] 3) Text feature enhancement extraction: Use adaptive text feature extraction algorithm to dynamically adjust feature extraction focus for different types of disputes; for example, contract disputes focus on clause analysis, and labor disputes focus on salary and working hour data extraction to improve text processing efficiency and accuracy.

[0114] II. Deep optimization of feature extraction: 1) Case relationship graph construction: Use Graph Neural Network (GNN) to extract graph structure features such as party relationship and event causal chain from dispute data, form case knowledge graph, and deeply mine dispute correlation information; for example, in group litigation, construct party relationship network and identify core characters and key event associations.

[0115] 2) Four-dimensional modeling of spatio-temporal features: Convert the case occurrence location into a three-dimensional geographic coordinate, combine the occurrence time to construct a spatio-temporal cube, extract spatio-temporal correlation features through convolutional neural network, and accurately locate the spatio-temporal pattern of disputes; for example, analyze regional consumer dispute hotspots and deploy mediation resources in advance.

[0116] 3) Social environment feature fusion: Introduce macroeconomic indicators, policy changes and other social environment data to build a social environment feature library; for example, combined with real estate policy adjustments, predict real estate dispute trends and optimize allocation strategies in advance.

[0117] Three, intelligent recognition model improvement: 1) Hybrid neural network architecture: Use LSTM to process text sequence features, and use CNN to extract case element spatial features, implement feature-level fusion through attention mechanism, and improve the accuracy of dispute type recognition; for example, in infringement disputes, integrate accident description text and scene picture elements to accurately determine the responsible party.

[0118] 2) Domain adaptive training: Design domain adaptive modules for different types of disputes to automatically adjust the attention of the model to specific domain features such as contract clauses and labor law clauses, enhancing the model's generalization ability; for example, the labor dispute model focuses on wage payment clauses, and the contract dispute model focuses on breach of contract clauses.

[0119] 3) Contrastive learning enhancement: Introduce contrastive learning mechanism to construct positive and negative sample pairs, and improve the model's ability to distinguish similar disputes; for example, distinguish between contract termination and contract termination disputes, and strengthen key feature recognition through contrastive learning.

[0120] Four, cross-department collaborative innovation mechanism: 1) Dynamic capability assessment system: Real-time acquisition of department processing efficiency (cases / hour), resource occupancy rate and other indicators, construction of dynamic department capability portrait, and realization of accurate case allocation; for example, predict the average time of complex cases handled by arbitration court based on historical data, and reasonably allocate new cases.

[0121] 2) Conflict resolution algorithm: When multiple departments claim jurisdiction at the same time, start the conflict resolution program based on semantic similarity and jurisdiction regulation knowledge graph, automatically determine the optimal receiving department, and avoid inter-departmental shirking; for example, when labor disputes involve infringement, assign to arbitration or mediation departments according to key feature weights.

[0122] 3) Collaborative processing flow optimization: design standardized collaborative processing procedures across departments, clearly define departmental responsibilities and interaction nodes, and improve overall processing efficiency; for example, mediation failure is automatically transferred to arbitration, and arbitration results are not satisfied and are guided to litigation, with seamless connection throughout the process.

[0123] After comparing with the related multi-resolution cross-departmental coordination method for resolving contradictions and disputes, the model accuracy in the above embodiment 1 is improved to 93.7% (compared with the related multi-resolution cross-departmental coordination method for resolving contradictions and disputes, which is improved by 18.3%), and in the labor dispute case, the accuracy is 96.2%, effectively reducing the workload of manual review; The average allocation time of the case is shortened to 1.2 seconds (the related multi-resolution cross-departmental coordination method for resolving contradictions and disputes needs 25.6 seconds), and the allocation efficiency is improved by 20 times in the peak period (such as the high incidence period of labor disputes at the beginning of the month); The delay of information transmission between departments is reduced by 82% (from an average of 4.3 hours to 24 minutes), and the coordination efficiency is significantly improved in the cross-regional intellectual property dispute processing; After the platform goes online, the regional contradiction and dispute resolution period is shortened by 34%, and the public satisfaction is improved to 91.2%, and in the family dispute processing, the settlement rate is improved to 93.6%.

[0124] The embodiment of the application also provides a multi-resolution cross-departmental coordination method for resolving contradictions and disputes, it should be noted that the steps shown in the flowchart of the drawing can be executed in a computer system such as a group of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from here.

[0125] In this embodiment, a multi-resolution cross-departmental coordination method for resolving contradictions and disputes is provided, which can be used in the multi-resolution cross-departmental coordination platform for resolving contradictions and disputes described above, Figure 1 is a flowchart of a multi-resolution cross-departmental coordination method for resolving contradictions and disputes according to the embodiment of the application, as Figure 5 shown, the flow includes the following steps: Step S501, the multi-source data acquisition and fusion module acquires multi-source heterogeneous contradiction and dispute data, and uses an adaptive fusion algorithm to fuse the multi-source heterogeneous contradiction and dispute data to obtain a contradiction and dispute description text.

[0126] Step S502, the feature extraction and intelligent identification module performs emotion analysis, time and space analysis, and environmental feature analysis on the contradiction and dispute description text to obtain a multi-dimensional contradiction and dispute feature space, and performs type identification based on the multi-dimensional contradiction and dispute feature space to obtain a contradiction and dispute type.

[0127] Step S503: The cross-departmental dynamic allocation module constructs a conflict resolution knowledge graph, uses the conflict resolution knowledge graph to determine the optimal conflict resolution handling department corresponding to the conflict type, and allocates conflict cases to the optimal conflict resolution handling department.

[0128] This embodiment presents a cross-departmental collaborative method for diversified dispute resolution, applied to, for example... Figure 1 The illustrated embodiment is a cross-departmental collaborative platform for diversified dispute resolution; therefore, the specific implementation methods of steps S501 and S503 can be referred to the preceding text. Figure 1 The corresponding descriptions of the illustrated embodiments are not repeated here.

[0129] It is understandable that the function and beneficial effects of the method in this embodiment are the same as those of the previous embodiment. Figure 1 The role and beneficial effects of the cross-departmental collaborative platform for diversified dispute resolution in the illustrated embodiment correspond to each other, and will not be elaborated here.

[0130] This invention also provides a computer device having the above-described features. Figure 1 This illustrates a cross-departmental collaborative platform for the diversified resolution of conflicts and disputes.

[0131] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 6 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 6 Take a processor 10 as an example.

[0132] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GPA), or any combination thereof.

[0133] The memory 20 stores instructions executable by the at least one processor 10 to cause the at least one processor 10 to perform the method illustrated in the above embodiments.

[0134] The memory 20 can include a program storage area and a data storage area. The program storage area can store an operating system and application programs required by at least one function. The data storage area can store data created according to the use of the computer device, and the like. In addition, the memory 20 can include a high-speed random access memory, and can also include a non-transitory memory such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some optional embodiments, the memory 20 can optionally include a memory disposed remotely with respect to the processor 10, and these remote memories can be connected to the computer device through a network. Examples of the above network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0135] The memory 20 can include a volatile memory such as a random access memory, and can also include a non-volatile memory such as a flash memory, a hard disk, or a solid-state disk. The memory 20 can also include a combination of the above-mentioned kinds of memories.

[0136] The computer device further includes a communication interface 30 for communication of the computer device with other devices or communication networks.

[0137] The embodiments of the present application also provide a computer readable storage medium. The above method according to the embodiments of the present application can be implemented in hardware, firmware, or recorded in a storage medium, or implemented as computer code originally stored in a remote storage medium or a non-transitory machine readable storage medium and downloaded to a local storage medium, so that the method described herein can be processed by such software on a storage medium using a general purpose computer, a special purpose processor, or programmable or special purpose hardware. The storage medium can be a magnetic disk, an optical disk, a read-only memory, a random access memory, a flash memory, a hard disk, or a solid-state disk, and the like. Further, the storage medium can also include a combination of the above-mentioned kinds of memories. It can be understood that the computer, the processor, the microprocessor controller, or the programmable hardware includes a storage component that can store or receive software or computer code, when the software or computer code is accessed and executed by the computer, the processor, or the hardware, the method illustrated in the above embodiments is implemented.

[0138] Part of the present application can be applied as a computer program product, for example, computer program instructions, when executed by a computer, through the operation of the computer, can invoke or provide the method and / or technical solutions according to the present application. Those skilled in the art should understand that the form of computer program instructions in computer readable medium includes but is not limited to source files, executable files, installation package files and the like, and accordingly, the way of computer program instructions executed by computer includes but is not limited to: the computer directly executes the instructions, or the computer compiles the instructions and then executes the corresponding compiled program, or the computer reads and executes the instructions, or the computer reads and installs the instructions and then executes the corresponding installed program. Here, the computer readable medium can be any available computer readable storage medium or communication medium accessible to the computer.

[0139] Although the embodiments of the present application are described in conjunction with the drawings, various modifications and changes can be made by those skilled in the art without departing from the spirit and scope of the present application, and such modifications and changes fall within the scope defined by the appended claims.

Claims

1. A multi-departmental collaborative platform for resolving contradictions, characterized in that, The system comprises: a data layer, a processing layer and a coordination layer connected in sequence; the data layer comprises a multi-source data acquisition and fusion module; the processing layer comprises a feature extraction and intelligent identification module; the coordination layer comprises a cross-department dynamic allocation module; the multi-source data acquisition and fusion module is configured to acquire multi-source heterogeneous contradictory dispute data, fuse the multi-source heterogeneous contradictory dispute data by using an adaptive fusion algorithm, and obtain contradictory dispute description text; the feature extraction and intelligent identification module is configured to perform emotion analysis, time-space analysis and environmental feature analysis on the contradictory dispute description text, obtain a multi-dimensional contradictory dispute feature space, and perform type identification based on the multi-dimensional contradictory dispute feature space to obtain a contradictory dispute type; the cross-department dynamic allocation module is configured to construct a conflict resolution knowledge graph, determine an optimal contradictory dispute handling department corresponding to the contradictory dispute type by using the conflict resolution knowledge graph, and allocate contradictory dispute cases to the optimal contradictory dispute handling department.

2. The multi-department collaborative platform for dispute resolution according to claim 1, characterized in that, The multi-source data acquisition and fusion module is specifically configured to acquire multi-source heterogeneous contradictory dispute data, perform data preprocessing on the multi-source heterogeneous contradictory dispute data, extract features from the data-preprocessed contradictory dispute data, obtain contradictory dispute features, dynamically adjust weights corresponding to the data-preprocessed contradictory dispute data, obtain feature weights, weight fuse the contradictory dispute features based on the feature weights, and obtain the contradictory dispute description text.

3. The multi-department collaborative platform for dispute resolution according to claim 1, characterized in that, The feature extraction and intelligent identification module comprises: an emotion tendency strengthening analysis unit configured to perform emotion identification and voice intonation identification on the contradictory dispute description text, obtain an emotion quantization level and a voice intonation feature, associate the emotion quantization level and the voice intonation feature, and obtain an emotion feature; a text feature enhancement extraction unit configured to extract contradictory dispute time-space features based on the contradictory dispute description text, obtain social environment features, fuse the emotion feature, the contradictory dispute time-space features and the social environment features, and obtain the multi-dimensional contradictory dispute feature space; a domain adaptive model training unit configured to obtain historical contradictory dispute case data, extract features from the historical contradictory dispute case data, determine contradictory dispute general features and contradictory dispute domain features, optimize and train a neural network model by using the historical contradictory dispute case data, the contradictory dispute general features and the contradictory dispute domain features, and obtain a domain adaptive model; a contradictory dispute type identification unit configured to input the multi-dimensional contradictory dispute feature space into the domain adaptive model, and obtain the contradictory dispute type.

4. The multi-department collaborative platform for dispute resolution according to claim 1, characterized in that, The cross-department dynamic allocation module comprises: a dynamic capability evaluation unit configured to obtain contradictory dispute handling indexes of a plurality of contradictory dispute handling departments, construct a dynamic department portrait based on the contradictory dispute handling indexes of the plurality of contradictory dispute handling departments; The conflict resolution unit is configured to construct a regulation knowledge graph, determine a regulation path corresponding to the type of the contradiction dispute by using the regulation knowledge graph, calculate semantic similarity based on a text of the type of the contradiction dispute and a regulation clause in the regulation path, and obtain a conflict resolution rule. The optimal contradiction dispute handling department is determined based on the semantic similarity, the conflict resolution rule, and the dynamic department portrait. The collaborative processing flow optimization unit is configured to establish a standardized cross-department collaborative processing framework, and allocate a contradiction dispute case by using the standardized cross-department collaborative processing framework based on the optimal contradiction dispute handling department.

5. The multi-department collaborative platform for dispute resolution according to claim 1, characterized in that, Further comprising: A security layer connected with the data layer, the processing layer, and the collaboration layer respectively; the security layer comprises a data encryption and access control module; the data encryption and access control module comprises a federated learning data collaboration framework construction unit and a multi-level encryption unit; The federated learning data collaboration framework construction unit is configured to obtain contradiction dispute handling departments participating in contradiction dispute processing, take the contradiction dispute handling departments participating in contradiction dispute processing as clients, and construct a federated learning data collaboration framework based on a central server and a plurality of clients. The central server is connected with a plurality of clients respectively; The multi-level encryption unit is configured to obtain local contradiction dispute data of a client, train and update a local contradiction dispute classification model based on the local contradiction dispute data of the client, obtain updated model parameters, encrypt the updated model parameters to obtain encryption parameters corresponding to the client, aggregate the encryption parameters by the central server to obtain global contradiction dispute classification model parameters, update a global contradiction dispute classification model by using the global contradiction dispute classification model parameters, encrypt the updated global contradiction dispute classification model, and send the encrypted global contradiction dispute classification model to the client for contradiction dispute classification processing.

6. The multi-department collaborative platform for dispute resolution according to claim 1, characterized in that, Further comprising: A display layer connected with the data layer, the processing layer, and the collaboration layer respectively; The display layer comprises a user interaction and visualization module; The user interaction and visualization module is configured to collect and store contradiction dispute processing data, and perform multi-terminal access control and custom configuration on the contradiction dispute processing data.

7. A multi-departmental conflict resolution and multi-resolution cross-departmental collaboration method, characterized in that, The method is applied to the contradiction dispute multi-resolution cross-department collaboration platform, and the method comprises the following steps: A multi-source data acquisition and fusion module acquires multi-source heterogeneous contradiction dispute data, and fuses the multi-source heterogeneous contradiction dispute data by using an adaptive fusion algorithm to obtain contradiction dispute description texts; A feature extraction and intelligent identification module performs emotion analysis, time-space analysis, and environmental feature analysis on the contradiction dispute description texts to obtain a multi-dimensional contradiction dispute feature space, and performs type identification based on the multi-dimensional contradiction dispute feature space to obtain a type of the contradiction dispute. The cross-department dynamic allocation module constructs a conflict resolution knowledge graph, determines an optimal dispute handling department corresponding to the type of the dispute based on the conflict resolution knowledge graph, and allocates the dispute case to the optimal dispute handling department.

8. A computer device, comprising: The method comprises the following steps: A memory and a processor are in communication connection with each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the multi-resolution cross-department coordination method for the dispute according to claim 7.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for causing a computer to execute the multi-resolution cross-department coordination method for the dispute according to claim 7.

10. A computer program product, characterised in that, The computer readable storage medium stores computer instructions for causing a computer to execute the multi-resolution cross-department coordination method for the dispute according to claim 7. The computer readable storage medium stores computer instructions for causing a computer to execute the multi-resolution cross-department coordination method for the dispute according to claim 7.

Images