Social governance event management method based on multi-source data fusion
By integrating multi-source data and applying AI models, the problems of data silos and dispatching errors in social governance event management have been solved, achieving accurate event allocation and process transparency, thereby improving governance efficiency and citizen satisfaction.
Patent Information
- Application Number
- CN202511940258.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-22
- Publication Date
- 2026-03-20
AI Technical Summary
The current social governance incident management suffers from problems such as fragmented multi-source data, low efficiency and insufficient accuracy of manual assignment, and lack of effective closed-loop and supervision mechanisms, resulting in information silos, assignment errors, and opaque processing procedures.
A multi-source data fusion method is adopted, high-dimensional feature vectors are generated through BERT/Word2Vec, and a weighted summation method is used to form unified event features. AI models are used for accurate classification and assignment, and a strategy for identifying and merging duplicate events is implemented to construct a closed-loop management and governance effectiveness quantitative assessment.
It has achieved improved integration of multi-source data, accurate event dispatch and load balancing, reduced invalid dispatches, shortened processing cycles, and improved governance efficiency, transparency and satisfaction.
Smart Images
Figure CN121707508A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of social governance and public service technology, specifically a social governance event management method based on multi-source data fusion. Background Technology
[0002] Social governance incident management refers to a systematic process in which multiple stakeholders, including the government, social organizations, market entities, and the public, use a series of systems, mechanisms, and technical means to prevent, warn against, handle, handle, and restore various incidents that affect social order, public safety, people's well-being, or social stability. Its core objectives are to resolve social conflicts, prevent and control social risks, maintain public order, protect public rights, and promote social harmony.
[0003] However, existing social governance practices suffer from the following problems due to the dispersed sources of incidents, cumbersome handling procedures, and inefficient cross-departmental collaboration:
[0004] Fragmented data from multiple sources creates information silos: Because event data comes from the 12345 hotline, grid patrol, petition system, etc., the inconsistent formats include not only text, but also forms and images, resulting in duplicate reporting of the same event and missing related information.
[0005] Manual assignment is inefficient and inaccurate: due to reliance on manual judgment, assignment errors and priority confusion are prone to occur, and the average assignment time exceeds 30 minutes;
[0006] Lack of effective closed-loop and supervision mechanisms: The incident handling process is not transparent and lacks scientific quantitative evaluation methods. Summary of the Invention
[0007] This invention provides a social governance event management method based on multi-source data fusion, which solves the problems mentioned in the background.
[0008] To achieve the above objectives, the present invention provides the following technical solution: a social governance event management method based on multi-source data fusion. The method comprises four modules: multi-source heterogeneous data fusion and feature standardization, an intelligent event allocation decision engine, a duplicate event identification and merging strategy, and full-process closed-loop management and governance effectiveness quantitative evaluation. The multi-source heterogeneous data fusion and feature standardization includes feature extraction and feature fusion. The intelligent event allocation decision engine includes an event classification and grading model and an intelligent allocation decision algorithm. The duplicate event identification and merging strategy employs a duplicate event identification algorithm. The full-process closed-loop management and governance effectiveness quantitative evaluation include full-process visual tracking and a governance effectiveness KPI quantification model.
[0009] Optionally, the feature extraction specifically refers to: using a BERT / Word2Vec pre-trained model to embed word vectors into the event description text to generate high-dimensional feature vectors. One-Hot encoding is used to generate structured data. Furthermore, structured data includes time, location, and type.
[0010] Optionally, the feature fusion specifically refers to: using a weighted summation method to form a unified event feature vector. ,and The calculation formulas are as follows:
[0011]
[0012] Where α, β, and γ are all empirical weights.
[0013] Optionally, the event intelligent dispatch decision engine uses AI models to accurately classify, rank, and dispatch events.
[0014] Optionally, the event classification and rating model specifically refers to a text classification model based on Bi-LSTM or Transformer architecture, for... The system is trained to output the event type and urgency level in real time, with the urgency level set to Level I, Level II, Level III, and Level IV.
[0015] The intelligent assignment decision algorithm specifically refers to: establishing a system that includes matching degrees of responsible departments. Departmental load and the urgency of the event The system selects the comprehensive allocation score function. The department with the highest value is assigned the task, and the assignment score function is used. The calculation formulas are as follows:
[0016]
[0017] Where λ, μ, and ƞ are all empirical weights.
[0018] Optionally, the duplicate event identification algorithm includes similarity calculation, merging determination, and merging strategy.
[0019] Optionally, the similarity calculation specifically refers to: calculating new events. Orders dispatched in the past 72 hours Overall similarity ,and The calculation formulas are as follows:
[0020]
[0021] Among them, w t w p w d These are the weighting coefficients for the three dimensions: time, location, and description, S. time S pos S desc Scores corresponding to time, location, and text dimensions, respectively, and text similarity. use Cosine similarity is used as a measure.
[0022] The decision point for the merging determination specifically refers to whether it is: ;
[0023] The merging strategy specifically refers to: when a duplicate is determined, [the strategy will be implemented]. Marked as The supervision list was improved. Priority.
[0024] Optionally, the full-process visual tracking specifically refers to: using the BPMN 2.0 model to construct an event handling process and recording timestamps for the five stages of acceptance, assignment, handling, feedback, and verification;
[0025] The governance effectiveness KPI quantification model includes evaluation indicators, specific indicator effectiveness scoring functions, and feedback loops.
[0026] Specifically, the evaluation indicators refer to: establishing an average processing cycle. Emergency response rate and citizen satisfaction A KPI evaluation system with [the following] as its core;
[0027] Performance scoring function The calculation formulas are as follows:
[0028]
[0029] The feedback loop specifically refers to the real-time feedback of the scoring results to the allocation decision engine of the second module, which serves as a correction item for the department's performance weight, thus forming a scheduling loop.
[0030] The present invention has the following beneficial effects:
[0031] 1. This social governance event management method based on multi-source data fusion adopts a heterogeneous event feature vector fusion method based on deep learning. Specifically, it uses BERT / Word2Vec to vectorize the text and combines a weighted summation method to fuse multi-source heterogeneous data into a unified event feature vector, avoiding information silos formed by fragmented multi-source data and improving the fusion of social governance event management data.
[0032] 2. This social governance event management method based on multi-source data fusion establishes a comprehensive allocation score function based on historical matching degree, real-time departmental load, and urgency by considering the intelligent allocation decision algorithm of departmental load and event urgency, thereby achieving accurate and load-balanced event allocation.
[0033] 3. This social governance event management method based on multi-source data fusion adopts a strategy for identifying and merging duplicate events based on spatial-temporal-textual similarity. By calculating the cosine similarity of event feature vectors and geographical distance, duplicate events are identified, and a strategy of sub-order supervision rather than duplicate order dispatch is adopted to effectively reduce invalid order dispatch.
[0034] 4. This social governance event management method based on multi-source data fusion adopts a closed-loop event handling process and efficiency quantification evaluation model based on core KPI indicators, constructs a BPMN process tracking system, and designs an efficiency scoring function with average processing cycle and satisfaction as the core to achieve real-time feedback from results to assignment decisions. Attached Figure Description
[0035] Figure 1 A flowchart for handling disputes related to property management in residential communities;
[0036] Figure 2 This is a general framework diagram of a social governance event intelligent dispatch system based on multi-source data fusion;
[0037] Figure 3 Flowchart for event feature vectorization and fusion;
[0038] Figure 4 Flowchart for duplicate event identification and merging strategy;
[0039] Figure 5 Flowchart of the intelligent dispatch decision engine;
[0040] Figure 6 This is a flowchart for closed-loop management and performance evaluation feedback of the entire event process. Detailed Implementation
[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0042] Please see Figure 1 and Figure 6This invention provides a technical solution: a social governance event management method based on multi-source data fusion. The method includes four modules, namely, multi-source heterogeneous data fusion and feature standardization, intelligent event allocation decision engine, duplicate event identification and merging strategy, and full-process closed-loop management and governance effectiveness quantitative evaluation. Multi-source heterogeneous data fusion and feature standardization includes feature extraction and feature fusion. The intelligent event allocation decision engine includes an event classification and grading model and an intelligent allocation decision algorithm. The duplicate event identification and merging strategy adopts a duplicate event identification algorithm. The full-process closed-loop management and governance effectiveness quantitative evaluation includes full-process visual tracking and a governance effectiveness KPI quantification model.
[0043] Feature extraction specifically refers to using a BERT / Word2Vec pre-trained model to embed word vectors into the event description text to generate high-dimensional feature vectors. One-Hot encoding is used to generate structured data. Furthermore, structured data includes time, location, and type.
[0044] Feature fusion specifically refers to using a weighted summation method to form a unified event feature vector. ,and The calculation formulas are as follows:
[0045]
[0046] Here, α, β, and γ are all empirical weights. The completeness of event information is improved after fusion. .
[0047] The event intelligent dispatch decision engine uses AI models to accurately classify, classify, and dispatch events.
[0048] Event classification and rating models specifically refer to text classification models based on Bi-LSTM or Transformer architectures, used for... The system is trained to output the event type and urgency level in real time, with the urgency level set to Level I, Level II, Level III, and Level IV.
[0049] The intelligent assignment decision algorithm specifically refers to: establishing a system that includes matching degree of responsible departments. Departmental load and the urgency of the event The system selects the comprehensive allocation score function. The department with the highest value is assigned the task, and the assignment score function is used. The calculation formulas are as follows:
[0050]
[0051] Where λ, μ, and ƞ are all empirical weights. The intelligent assignment accuracy reaches... The average distribution time has been reduced to minute.
[0052] Duplicate event identification algorithms include similarity calculation, merging determination, and merging strategy.
[0053] Similarity calculation specifically refers to: calculating new events Orders dispatched in the past 72 hours Overall similarity ,and The calculation formulas are as follows:
[0054]
[0055] Among them, w t w p w d These are the weighting coefficients for the three dimensions: time, location, and description, S. time S pos S desc Scores corresponding to time, location, and text dimensions, respectively, and text similarity. use Cosine similarity is used as a measure.
[0056] The decision point for merging refers specifically to whether it is: ;
[0057] The merging strategy specifically refers to: when a duplicate is determined, [the strategy will be implemented]. Marked as The supervision list was improved. Priority. The rate of identifying duplicate events has been increased to... .
[0058] Full-process visual tracking specifically refers to: using the BPMN 2.0 model to construct an event handling process and recording timestamps for the five stages of acceptance, assignment, handling, feedback, and verification;
[0059] The governance effectiveness KPI quantification model includes evaluation indicators, specific performance scoring functions, and feedback loops;
[0060] Specifically, the evaluation indicators refer to: establishing an average processing cycle. Emergency response rate and citizen satisfaction A KPI evaluation system with [the following] as its core;
[0061] Performance scoring function The calculation formulas are as follows:
[0062]
[0063] The feedback loop specifically refers to the real-time feedback of scoring results to the allocation decision engine in the second module, serving as a correction factor for departmental performance weights, thus forming a scheduling loop. The average processing time has been shortened to 36 hours, and governance satisfaction has increased. .
[0064] Example:
[0065] Handling disputes related to property management in residential communities:
[0066] S1, Data Input and Fusion
[0067] 1.1 Structured
[0068] 12345 hotline record: "A resident on the 5th floor of building 3 in this community complained that the property management fees were unreasonable." );
[0069] 1.2 Text
[0070] Grid worker's patrol record: "A protest letter was found posted at the entrance of the 5th floor of Building 3, involving a property dispute." );
[0071] The first module processes the text ("property disputes", "protest letters") using BERT vectorization and then weightedly merges it with structured data (addresses) to generate a unified event vector. ;
[0072] S2, Repeat Event Recognition
[0073] 2.1 Three-module processing: Compare with historical events within 48 hours Similarity calculation;
[0074] 2.2 Judgment Result: Judgment (higher than) If the threshold is reached, the system identifies it as a duplicate event, merges it with "Property Management Complaints" that were dispatched 24 hours ago, does not dispatch new orders, and raises the priority of the old orders by one level.
[0075] S3, Intelligent Allocation Decision
[0076] 3.1 Model Classification: Based on The classification model outputs: "Event Type: Property Management Dispute", "Importance Level: Level III (General)".
[0077] 3.2 Assignment Calculation: For this type of event, the system calculates the assignment for each department. :
[0078]
[0079]
[0080] 3.3 Assignment Result: The system automatically assigns the task to [the appropriate department / organization] within 5 seconds. The highest-level housing and construction department avoided the possibility of human resources being mistakenly assigned to the urban management department;
[0081] S4, Closed Loop and Evaluation
[0082] 4.1 Fourth Module Tracking: The process enters the handling stage. The housing and construction department completes mediation within 30 hours, and the system records the processing cycle. Hours (below average) Hour);
[0083] 4.2 Performance Evaluation: Improvement, calculation of the housing and construction department The score increases, and this score is fed back to the second module in real time, serving as a basis for the department's future performance. The implicit weights are adjusted in the calculation.
[0084] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. The terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, unless otherwise explicitly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. Moreover, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.
[0085] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A social governance event management method based on multi-source data fusion, characterized by: The method comprises four modules, namely, the first module, the second module, the third module, and the fourth module, which are respectively: multi-source heterogeneous data fusion and feature standardization, intelligent event dispatch decision engine, duplicate event identification and merging strategy, and full-process closed-loop management and governance effectiveness quantitative evaluation. The multi-source heterogeneous data fusion and feature standardization includes feature extraction and feature fusion. The intelligent event dispatch decision engine includes an event classification and grading model and an intelligent dispatch decision algorithm. The duplicate event identification and merging strategy adopts a duplicate event identification algorithm. The full-process closed-loop management and governance effectiveness quantitative evaluation includes full-process visual tracking and a governance effectiveness KPI quantification model.
2. The social governance event management method based on multi-source data fusion according to claim 1, characterized in that: The feature extraction specifically refers to: using a BERT / Word2Vec pre-trained model to embed word vectors into the event description text to generate high-dimensional feature vectors. One-Hot encoding is used to generate structured data. Furthermore, structured data includes time, location, and type.
3. The social governance event management method based on multi-source data fusion according to claim 1, characterized in that: The feature fusion specifically refers to: using a weighted summation method to form a unified event feature vector. ,and The calculation formulas are as follows: Where α, β, and γ are all empirical weights.
4. The social governance event management method based on multi-source data fusion according to claim 1, characterized in that: The event intelligent dispatch decision engine uses AI models to accurately classify, classify, and dispatch events.
5. The social governance event management method based on multi-source data fusion according to claim 1, characterized in that: The event classification and rating model specifically refers to a text classification model based on Bi-LSTM or Transformer architecture, used for... The system is trained to output the event type and urgency level in real time, with the urgency level set to Level I, Level II, Level III, and Level IV. The intelligent assignment decision algorithm specifically refers to: establishing a system that includes matching degrees of responsible departments. Departmental load and the urgency of the event The system selects the comprehensive allocation score function. The department with the highest value is assigned the task, and the assignment score function is used. The calculation formulas are as follows: Where λ, μ, and ƞ are all empirical weights.
6. The social governance event management method based on multi-source data fusion according to claim 1, characterized in that: The duplicate event identification algorithm includes similarity calculation, merging determination, and merging strategy.
7. The social governance event management method based on multi-source data fusion according to claim 6, characterized in that: The similarity calculation specifically refers to: calculating new events. Orders dispatched in the past 72 hours Overall similarity ,and The calculation formulas are as follows: Among them, w t w p w d These are the weighting coefficients for the three dimensions: time, location, and description, S. time S pos S desc Scores corresponding to time, location, and text dimensions, respectively, and text similarity. use Cosine similarity is used as a measure. The decision point for the merging determination specifically refers to whether it is: ; The merging strategy specifically refers to: when a duplicate is determined, [the strategy will be implemented]. Marked as The supervision list was improved. Priority.
8. The social governance event management method based on multi-source data fusion according to claim 1, characterized in that: The aforementioned full-process visualization tracking specifically refers to: using the BPMN 2.0 model to construct an event handling process, and recording timestamps for the five stages of acceptance, assignment, handling, feedback, and verification; The governance effectiveness KPI quantification model includes evaluation indicators, specific indicator effectiveness scoring functions, and feedback loops. Specifically, the evaluation indicators refer to: establishing an average processing cycle. Emergency response rate and citizen satisfaction A KPI evaluation system with [the following] as its core; Performance scoring function The calculation formulas are as follows: The feedback loop specifically refers to the real-time feedback of the scoring results to the allocation decision engine of the second module, which serves as a correction item for the department's performance weight, thus forming a scheduling loop.