Intelligent recommendation method and device for grassroots social grid complex event governance mechanism
The method and device enhance the accuracy and efficiency of complex event governance in grassroots social grid governance by using a simple-complex classification model and large language models to analyze event semantics, improving the handling of complex events.
Patent Information
- Application Number
- CN202411206209.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-08-30
AI Technical Summary
Current intelligent recommendation technologies for complex events in grassroots social grid governance lack accuracy and efficiency, relying heavily on deep learning models for classification and prediction, which struggle to handle complex event relationships and elements effectively.
A method and device that utilize a simple-complex classification model to differentiate between simple and complex events, combining vectorized retrieval and machine learning to recommend processing mechanisms, leveraging a large language model for semantic analysis and feature extraction to enhance accuracy.
The method and device improve the accuracy and efficiency of complex event governance by rapidly responding to and effectively handling complex events through intelligent mechanism recommendations, integrating expert experience and machine learning models.
Smart Images

Figure CN119088947A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of digital social governance, and in particular relates to an intelligent recommendation method and device for complex event mechanisms in grassroots social grid governance. Background Art
[0002] At present, there is no authoritative and scientific way and method for handling complex incidents in the grassroots social grid governance work. Mediation is mainly carried out through means such as source of conflict management and multi-party consultation. It relies more on the level of economic development and expert experience. There are problems such as unclear responsibilities between horizontal organizations and uncoordinated vertical relationships. There is a lack of scientific and standardized guiding norms for grassroots social grid governance.
[0003] It has become an industry consensus that deep learning models have good results in the field of natural language processing. In the complex events of grassroots social grid governance, the current mainstream mechanism intelligent recommendation technology is based on natural language models such as Bert and Transformer, which mainly encode and classify text output.
[0004] In the actual practice, it was found that the current intelligent mechanism recommendation technology, when dealing with complex mechanism relationships and event elements, especially in the recommendation of mechanisms for complex events, directly relies on deep learning models for classification and prediction, which is difficult to achieve accuracy and efficiency. Summary of the invention
[0005] In view of the problems existing in the prior art, the purpose of the embodiments of the present application is to provide a method and device for intelligently recommending a mechanism for complex event governance in grassroots social grids.
[0006] According to a first aspect of an embodiment of the present application, a method for intelligently recommending a mechanism for complex event governance in a grassroots social grid is provided, comprising: Connecting to a third-party information platform, obtaining pending events of grassroots social grid governance through an interface or direct database connection, and extracting a first feature from the pending events; Using a simple-complex classification model, based on the first feature, the event to be processed is classified into a simple event or a complex event; If the event to be processed is a complex event, a corresponding processing mechanism is recommended for it through vectorized retrieval and machine learning methods, and the first probability and second probability of belonging to each processing mechanism are obtained respectively. The first probability and second probability are sorted, so as to recommend the processing mechanism with the highest probability.
[0007] Furthermore, extracting the first feature of the event to be processed is specifically: If the event to be processed is structured data, directly extract the first feature; If the event to be processed is unstructured data, the first feature is extracted by manual annotation and element extraction.
[0008] Furthermore, a corresponding processing mechanism is recommended for the event to be processed by a vectorized retrieval method, and a first probability of belonging to each processing mechanism is obtained respectively, including: Performing vectorized encoding on the text content of the event to be processed; The encoded events are searched in the Elasticsearch database storing event cases to obtain several similar events with the highest similarity. According to the processing mechanisms corresponding to the similar events, a first probability that the event to be processed belongs to each processing mechanism is set.
[0009] Furthermore, a corresponding processing mechanism is recommended for the event to be processed by a machine learning method, and a second probability of belonging to each processing mechanism is obtained respectively, including: Reshaping the semantics of the event to be processed; Using a large language model, extract keywords from the semantically reshaped events to be processed and quantify them to obtain the event degree quantification value; Extracting the second feature of the event to be processed, wherein the second feature includes the number of persons involved, the quantitative value of the subject amount, the quantitative value of the event degree, the event type and the event theme; Based on the second feature, the event to be processed is classified using a pre-trained multi-classification model to obtain a second probability that the event belongs to each processing mechanism.
[0010] Furthermore, the large language model is used to extract and quantify the keywords of the semantically reshaped events to be processed, and obtain the quantitative value of the event degree, which is specifically: Use a large language model to extract keywords from the semantically reshaped events to be processed; According to the semantic fuzziness of the keyword and the influence of the corresponding behavior, the keyword is quantified to obtain a quantified value of the event degree.
[0011] Furthermore, the quantified value of the target amount is obtained in the following manner: The maximum amount among the extracted amounts of the pending events is taken as the target amount; The target amount is matched to a corresponding quantified value according to a predetermined segmentation range, thereby obtaining a quantified value of the target amount.
[0012] According to a second aspect of the embodiment of the present application, there is provided a grassroots social grid governance complex event mechanism intelligent recommendation device, comprising: A feature extraction module is used to connect to a third-party information platform, obtain pending events of grassroots social grid governance by means of an interface or a direct database connection, and extract a first feature from the pending events; A traditional-simplified classification module, configured to classify the event to be processed into a simple event or a complex event based on the first feature by using a traditional-simplified classification model; The mechanism recommendation module is used to recommend a corresponding processing mechanism for the event to be processed through vectorized retrieval and machine learning methods if the event to be processed is a complex event, obtain the first probability and the second probability of belonging to each processing mechanism respectively, sort the first probability and the second probability, and thus recommend the processing mechanism with the highest probability.
[0013] According to a third aspect of an embodiment of the present application, a computer program product is provided, comprising a computer program / instruction. When the computer program / instruction is executed by a processor, the method described in the first aspect is implemented.
[0014] According to a fourth aspect of an embodiment of the present application, there is provided an electronic device, including: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in the first aspect.
[0015] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the method described in the first aspect are implemented.
[0016] The technical solution provided by the embodiments of the present application may have the following beneficial effects: It can be seen from the above embodiments that the present application constructs a method and device for intelligently recommending mechanisms for complex event governance at the grassroots social grid level in the field of social governance. The present application utilizes event data reported by grassroots staff or connected to third-party software platforms, and fully combines expert experience and machine learning related models to divide them into simple events and complex events. Then, for complex events, the method of vectorized retrieval and machine learning is combined to make intelligent recommendation of mechanisms, and the deep neural network model and the large prediction model are used to fully analyze the intrinsic semantics, elements and other factors of complex events, and the prediction of processing mechanisms is realized, which can achieve rapid response and effective disposal of complex events. The use of intelligent technology can further improve the accuracy and efficiency of governance, and inject new vitality into grassroots social governance.
[0017] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.
[0019] Figure 1 It is a flow chart of an intelligent recommendation method for a complex event mechanism for grassroots social grid governance according to an exemplary embodiment.
[0020] Figure 2 is a flowchart of step S3 according to an exemplary embodiment.
[0021] Figure 3 It is a block diagram of an intelligent recommendation device for a complex event mechanism for grassroots social grid governance according to an exemplary embodiment.
[0022] Figure 4 The diagram is a schematic diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION
[0023] Here, exemplary embodiments are described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application.
[0024] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.
[0025] It should be understood that although the terms first, second, third, etc. may be used in the present application to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".
[0026] Figure 1 is a flow chart of an intelligent recommendation method for complex event mechanism of grassroots social grid governance according to an exemplary embodiment. Figure 1 As shown, the method may include the following steps: S1: Connecting to a third-party information platform, obtaining pending events of grassroots social grid governance through an interface or direct database connection, and extracting a first feature from the pending events; S2: using a simple-complex classification model to classify the event to be processed into a simple event or a complex event based on the first feature; S3: If the event to be processed is a complex event, a corresponding processing mechanism is recommended for it through vectorized retrieval and machine learning methods, and the first probability and second probability of belonging to each processing mechanism are obtained respectively, and the first probability and second probability are sorted to recommend the processing mechanism with the highest probability.
[0027] It can be seen from the above implementation that this application constructs a method and device for intelligent recommendation of complex event mechanisms for grassroots social grid governance in the field of social governance. This application uses event data reported by grassroots staff or connected to third-party software platforms, and fully combines expert experience and machine learning related models to divide them into simple events and complex events. Then, for complex events, the method of vectorized retrieval and machine learning is combined to make intelligent recommendation of mechanisms, and the deep neural network model and the large prediction model are used to fully analyze the inherent semantics, elements and other factors of complex events to realize the prediction of processing mechanisms, which can achieve rapid response and effective disposal of complex events. The use of intelligent technology can further improve the accuracy and efficiency of governance and inject new vitality into grassroots social governance.
[0028] In a specific implementation of S1, a third-party information platform is connected to obtain pending events of grassroots social grid governance through an interface or a direct database connection, and a first feature is extracted from the pending events; Specifically, in the field of social governance, there are different information system platforms in different regions, and their data structures are similar, including basic information of users, case content, and reporting time, etc., which can be reported by staff or by users themselves; manual annotation is performed in combination with expert experience, and the reported events are structured through methods such as element extraction to extract the first feature. Among them, the first feature includes: the number of people involved in the case, the amount of the subject matter, the situation of the people involved, the aggravated behavior, the potential violence, and the degree of urgency. In one embodiment, it can be set as follows: Number of people involved: at least 1 person. If the same incident involves multiple people, there is a possibility of a collective incident; Target amount: The minimum target amount is 0 yuan. Generally, the larger the target amount, the higher the severity of the event; Situation of the person involved in the case: There are two situations, namely, death occurred and no death occurred; Agitation behavior: Based on expert experience, weights are assigned to some sensitive words (such as arrears, strike, etc.) contained in the event content to form an aggravation behavior form.
[0029] Potential violence: Based on expert experience, potential violence (such as quarrels, fractures, traffic accidents, etc.) is divided into 9 levels and a form is constructed.
[0030] Urgency: Based on expert experience, incidents are divided into four levels: major and urgent, major but not urgent, urgent but not major, and non-major and urgent.
[0031] That is, based on six indicators representing event characteristics, the unstructured data of the grassroots social grid governance events are processed through manual annotation mode and factor extraction technology to form structured data. Manual annotation includes assessments of potential violence, radicalized behavior, degree of urgency, etc.
[0032] It should be noted that the acquired events may also be data that has been structured according to the above requirements, and no further structuring is required.
[0033] In a specific implementation of S2, a complex-simplified classification model is adopted to classify the event to be processed into a simple event or a complex event based on the first feature; Specifically, the structured data is classified into two categories through a pre-trained machine learning model (logistic regression model), and divided into simple events or complex events, as shown in Table 1: Table 1 Examples of simple and complex events content Label I live in xx Street, xx City. Recently, I need to apply for a temporary residence permit. A policeman from xx community contacted me and asked me to go to the community to apply for a temporary residence permit. I have questions about this. I would like to inquire about where I should go to apply for a temporary residence permit and whether there is any charge. Please tell me the specific process. Simple Events The xx Cement Products Factory near xx Village, xx Street, xx City is currently producing stone powder. When the wind is strong, all the dust is blown into the villagers' houses. The environmental protection department has previously required it to rectify the situation, but it has not yet carried out an environmental protection publicity and is still producing stone powder. Please ask the department to verify and rectify the situation. Complex events If it is a simple event, no further processing is performed; if it is a complex event, the processing of step S3 is continued.
[0034] In a specific implementation of S3, if the event to be processed is a complex event, a corresponding processing mechanism is recommended for it through vectorized retrieval and machine learning methods, and the first probability and second probability of belonging to each processing mechanism are obtained respectively, and the first probability and second probability are sorted, so as to recommend the processing mechanism with the highest probability. Figure 2 shown.
[0035] Specifically, this application involves three processing mechanisms, namely the first mechanism, the second mechanism and the third mechanism. The operation of the first mechanism is the responsibility or leadership of the township level; the operation of the second mechanism is the responsibility or leadership of the village and community level; the operation of the third mechanism is the responsibility or leadership of the district and county level. Specifically, the first mechanism is for complex incidents at the township level. In the process of resolving the incidents, it is necessary to involve the collaboration of multiple departments, and adopt a multi-faceted collaborative approach to circulate and handle matters; the second mechanism is for complex incidents at the village and community level. The incidents involve collective interests such as road and bridge construction, which require multiple people to negotiate together, and the village and community cadres negotiate with the masses to achieve resolution and resolution; the third mechanism is for complex incidents at the district and county level, which are usually sudden emergencies or major hazards, and the relevant responsible persons and experts are convened to discuss and judge them, and a handling decision is made.
[0036] Specifically, recommending a corresponding processing mechanism for the event to be processed by a vectorized retrieval method, and obtaining the first probability of the event belonging to each processing mechanism respectively, may include: S11: performing vectorized encoding on the text content of the event to be processed; This application uses the M3E model series for text vectorization encoding. Since the events involved in this application are usually short text data, Roberta's basic model is used for encoding. The input length is about 512 tokens, which is sufficient for the text encoding tasks of most events. If it exceeds the input length, it will be truncated. The output is a 768-dimensional vector, that is, no matter how long the input text is, the output vector length is fixed. That is, a 768-dimensional vector is used to represent the event; S12: searching the encoded event in an Elasticsearch database storing event cases to obtain several similar events with the highest similarity; In practical applications, as the case base grows, challenges arise in expanding the knowledge base vector and ensuring efficient search capabilities. To solve this problem, Elasticsearch software is introduced to implement the knowledge base addition, deletion, modification, query, and vectorized retrieval or keyword retrieval.
[0037] The 768-dimensional vectors are stored in the Elasticsearch database, and cosine similarity is used as a metric to compare the similarity between events. Cosine similarity evaluates the similarity between two non-zero vectors by measuring the cosine value of the angle between them. Its value ranges from -1 to 1, where 1 represents the exact same direction, 0 represents orthogonal, and -1 represents the exact opposite direction. In text retrieval, cosine similarity is often used for the similarity between text vectors.
[0038] In Elasticsearch, vectorized retrieval of cosine values is implemented through the following steps: S121: Create an index in Elasticsearch and define a dense_vector type field to store the vector representation of the event, and store the encoded text in the dense_vector field; S123: using the vector search function of Elasticsearch to search for events with the most similar vectors. In a specific implementation, the cosine similarity between the query vector and each event vector in the index may be calculated, and the results may be sorted according to the similarity scores. S124: When the results are returned, Elasticsearch will return a list of events sorted by similarity scores. These events are most similar to the query vector and have corresponding similarity scores. The TopK events are set to be returned.
[0039] S13: according to the processing mechanisms corresponding to the similar events, setting the first probability that the event to be processed belongs to each processing mechanism; Specifically, according to the processing mechanism corresponding to each similar event, the first probability that the event to be processed belongs to each processing mechanism is set. For example, 5 similar events are retrieved, of which three events are processed using the third mechanism and two events are processed using the second mechanism. Then, the first probability of the third mechanism is set to 60%, the first probability of the second mechanism is set to 40%, and the first probability of the first mechanism is set to 0.
[0040] In a specific implementation, the first probability can also be set in combination with the similarity between each similar event and the event to be processed. For example, 5 similar events are retrieved, of which events 1, 2, and 4 (similarities are 95%, 90%, and 80%, respectively) are processed using the third mechanism, and events 3 and 5 are processed using the second mechanism (similarities are 85% and 60%, respectively). Then, the first probability of the third mechanism is set to (95%+90%+80%) / 3*60%=54%, the first probability of the second mechanism is set to (85%+60%) / 2*40%=29%, and the first probability of the first mechanism is 0. It should be noted that only an example of setting the first probability in combination with similarity is given here, and other combinations can also be used in specific implementations.
[0041] The method of obtaining probability through vectorized retrieval is simple and convenient, but sometimes the accuracy may not be high and the recognition ability is limited, such as the situation where the similarity between the retrieved similar events and the event to be processed is low. Therefore, in order to achieve more accurate case classification, for cases that cannot be directly matched using simple retrieval technology, the present invention uses a method combining large language model (LLM) with machine learning, so that the events can be successfully classified.
[0042] Specifically, recommending a corresponding processing mechanism for the event to be processed by a machine learning method, and obtaining the second probability of the event belonging to each processing mechanism respectively, may include: S21. Reshaping the semantics of the event to be processed; Specifically, this step belongs to event preprocessing. Events have common problems such as colloquialism and typos. The large language model can achieve the function of semantic reshaping for such situations, thereby improving event quality.
[0043] Specifically, for semantically ambiguous event texts, feature extraction cannot effectively obtain key information, which may still affect the performance of classification results. For this type of text, the present invention further uses the semantic reshaping function in the big model technology to reprocess the text. In the first stage, a big language model is used to rewrite the text to refine the case description while retaining its semantic content. This method ensures the accuracy of structure and semantics, thereby enhancing the semantic interpretation ability of the traditional classification model. Since the big language model performs well in sentence reorganization in the analysis of legal texts. As a generative model trained on a wide range of text corpora, the big language model has strong generalization ability in tasks related to text rewriting. Rewriting the case text using the big language model not only retains the original meaning of the case text, but also improves clarity and coherence. Using the semantic understanding of the big language model, relevant keywords are placed in appropriate positions in the case text, improving their readability. In this rewriting stage, the big language model identifies key semantic elements and seamlessly integrates them into the revised text, thereby enriching the text narrative. For subsequent classification tasks, accurate classification requires in-depth semantic understanding. Traditional machine learning methods rely heavily on feature engineering and may not provide the depth of semantic analysis required for accurate classification. Therefore, this application turns to a neural network model for case classification, integrating the text rewriting and semantic interpretation capabilities of a large language model, which greatly enhances the ability of the neural network model to process and classify complex legal texts, thereby ensuring data readiness for machine learning tasks and improving the efficiency of the model.
[0044] S22. Using the large language model, extract and quantify keywords from the semantically reshaped events to be processed to obtain a quantitative value of the event degree; For example, the first mechanism is used to deal with a complex event involving the participation of multiple departments; the second mechanism is used to deal with a complex event involving the common interests of the people in the same village community, such as bridge repair or community reform; the third mechanism is mainly used to deal with major emergencies, such as landslides, urban flooding and other complex events. Starting from the differences between the three mechanisms, combined with expert experience and large models, feature extraction and manual labeling of complex events are achieved, and a supervised machine learning data set is constructed. The label of each complex event is one of the three mechanisms.
[0045] The present invention can conduct a comprehensive analysis of the frequency of words in the text in the training set. Several feature groups are established to refine the keyword set, and several representative keywords are determined to fill these sets based on statistical results and expert insights. Compared with traditional methods, the large language model has excellent capabilities in named entity recognition (NER), so the large language model is used as a feature extraction tool to extract key features. The large language model has a deep understanding of the overall semantic logic of sentences, keywords, and text. The large language model also effectively handles the ambiguity of the case text; it can discern the semantics of keywords even if there is no clear keyword indication in the sentence. In the keyword extraction process, the situation where the keywords are clearly present in the text is classified as semantically accurate. In contrast, if the text implies the meaning of the keywords without directly mentioning them, it is considered to be semantically ambiguous. After extracting the keywords, the case is evaluated using the feature groups constructed in the above steps, and cases of different complexity are scored.
[0046] According to the advice of legal experts and the importance of certain keywords, for keyword-sensitive words such as "house demolition", "law", and "traffic accident", no matter they are semantically precise or vague, as long as the large language model successfully extracts them, the corresponding features will get 1 point. In particular, for "law", as long as any word in the keyword group exists, 1 point can be obtained. For "domestic violence" and "threat", if they are semantically accurate, the score is 1; if they are semantically ambiguous, 0.5 points are assigned. For feature groups 1 and 2, according to the degree of adverse effects caused by their respective behaviors, when the semantics are precise, scores of 0.5 and 0.2 are assigned, and when the semantics are ambiguous, half the score is assigned, so that a quantitative value of the degree of the event can be obtained.
[0047] S23: extracting the second feature of the event to be processed, wherein the second feature includes the number of persons involved, the quantitative value of the target amount, the quantitative value of the event degree, the event type and the event theme; Specifically, the amount described in the event is also important for determining the complexity of the case. Therefore, a "quantity" function group is introduced. The maximum values of various quantities in the text are extracted to calculate the features, and the quantity is divided into six intervals using the box segmentation method, which are set to: [0, 1000), [1000, 10000), [10000, 50000), [50000, 100000) and [100000, +∞) in one embodiment. Each interval corresponds to a different score range of 0 to 5. The event theme is screened from the keywords extracted from the event content, and the event type is a manually annotated multi-level event label.
[0048] S23. Based on the second feature, the event to be processed is classified using a pre-trained multi-classification model to obtain a second probability that the event belongs to each processing mechanism; Specifically, a traditional machine learning model logistic regression algorithm is used, which is a three-classification model in this embodiment, and is trained using the second feature of the event and the annotated processing mechanism as labels.
[0049] After obtaining the first probability and the second probability that the event to be processed belongs to each mechanism, the first probability and the second probability are sorted, and the mechanism with the highest probability value is determined as the recommended mechanism for the event to be processed.
[0050] Corresponding to the aforementioned embodiment of the method for intelligently recommending a mechanism for complex events in grassroots social grid governance, the present application also provides an embodiment of an apparatus for intelligently recommending a mechanism for complex events in grassroots social grid governance.
[0051] Figure 3 This is a block diagram of an intelligent recommendation device for complex event mechanism of grassroots social grid governance according to an exemplary embodiment. Figure 3 , the device may include: The feature extraction module 21 is used to connect to a third-party information platform, obtain pending events of grassroots social grid governance by means of an interface or a direct database connection, and extract a first feature from the pending events; A traditional-simplified classification module 22, configured to classify the to-be-processed event into a simple event or a complex event based on the first feature by using a traditional-simplified classification model; The mechanism recommendation module 23 is used to recommend a corresponding processing mechanism for the event to be processed through vectorized retrieval and machine learning methods if the event to be processed is a complex event, obtain the first probability and the second probability of belonging to each processing mechanism respectively, sort the first probability and the second probability, and thus recommend the processing mechanism with the highest probability.
[0052] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0053] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present application scheme. A person of ordinary skill in the art can understand and implement it without paying any creative work.
[0054] Correspondingly, the present application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the above-mentioned intelligent recommendation method for complex event mechanisms in grassroots social grid governance.
[0055] Accordingly, the present application also provides an electronic device, including: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned grassroots social grid governance complex event mechanism intelligent recommendation method. Figure 4 As shown, it is a hardware structure diagram of any device with data processing capability where a grassroots social grid governance complex event mechanism intelligent recommendation device provided by an embodiment of the present invention is located, except Figure 4 In addition to the processor, memory and network interface shown, any device with data processing capability in which the apparatus in the embodiment is located may also include other hardware, generally based on the actual functions of the device with data processing capability, which will not be described in detail.
[0056] Accordingly, the present application also provides a computer-readable storage medium on which computer instructions are stored, and when the instructions are executed by the processor, the above-mentioned intelligent recommendation method for the complex event mechanism of grassroots social grid governance is implemented. The computer-readable storage medium can be an internal storage unit of any device with data processing capabilities described in any of the aforementioned embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), an SD card, a flash card (Flash Card), etc. equipped on the device. Furthermore, the computer-readable storage medium can also include both an internal storage unit and an external storage device of any device with data processing capabilities. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing capabilities, and can also be used to temporarily store data that has been output or is to be output.
[0057] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the contents disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not disclosed in the present application.
[0058] It should be understood that various modifications and changes can be made to the present application without departing from the scope of what has been described above and shown in the accompanying drawings.
Claims
1. An intelligent recommendation method for complex event mechanism of grassroots social grid governance, characterized by: include: Connecting to a third-party information platform, obtaining pending events of grassroots social grid governance through an interface or direct database connection, and extracting a first feature from the pending events; Using a simple-complex classification model, based on the first feature, the event to be processed is classified into a simple event or a complex event; If the event to be processed is a complex event, a corresponding processing mechanism is recommended for it through vectorized retrieval and machine learning methods, and the first probability and second probability of belonging to each processing mechanism are obtained respectively. The first probability and second probability are sorted, so as to recommend the processing mechanism with the highest probability.
2. The method according to claim 1, characterized in that: Extracting the first feature of the event to be processed is specifically: If the event to be processed is structured data, directly extract the first feature; If the event to be processed is unstructured data, the first feature is extracted by manual annotation and element extraction.
3. The method according to claim 1, characterized in that Recommending a corresponding processing mechanism for the event to be processed by a vectorized retrieval method, and obtaining a first probability of the event belonging to each processing mechanism, respectively, including: Performing vectorized encoding on the text content of the event to be processed; The encoded events are searched in the Elasticsearch database storing event cases to obtain several similar events with the highest similarity. According to the processing mechanisms corresponding to the similar events, a first probability that the event to be processed belongs to each processing mechanism is set.
4. The method according to claim 1, characterized in that: Recommending a corresponding processing mechanism for the event to be processed by a machine learning method, and obtaining a second probability of the event belonging to each processing mechanism, respectively, including: Reshaping the semantics of the event to be processed; Using a large language model, extract keywords from the semantically reshaped events to be processed and quantify them to obtain the event degree quantification value; Extracting the second feature of the event to be processed, wherein the second feature includes the number of persons involved, the quantitative value of the subject amount, the quantitative value of the event degree, the event type and the event theme; Based on the second feature, the event to be processed is classified using a pre-trained multi-classification model to obtain a second probability that the event belongs to each processing mechanism.
5. The method according to claim 4, characterized in that Using the large language model, we extract and quantify the keywords of the semantically reshaped events to be processed, and obtain the quantitative value of the event degree, which is as follows: Use a large language model to extract keywords from the semantically reshaped events to be processed; According to the semantic fuzziness of the keyword and the influence of the corresponding behavior, the keyword is quantified to obtain a quantified value of the event degree.
6. The method according to claim 4, characterized in that The quantified value of the target amount is obtained in the following way: The maximum amount among the extracted amounts of the pending events is taken as the target amount; The target amount is matched to a corresponding quantified value according to a predetermined segmentation range, thereby obtaining a quantified value of the target amount.
7. An intelligent recommendation device for complex event mechanism of grassroots social grid governance, characterized in that: include: A feature extraction module is used to connect to a third-party information platform, obtain pending events of grassroots social grid governance by means of an interface or a direct database connection, and extract a first feature from the pending events; A traditional-simplified classification module, configured to classify the event to be processed into a simple event or a complex event based on the first feature by using a traditional-simplified classification model; The mechanism recommendation module is used to recommend a corresponding processing mechanism for the event to be processed through vectorized retrieval and machine learning methods if the event to be processed is a complex event, obtain the first probability and the second probability of belonging to each processing mechanism respectively, sort the first probability and the second probability, and thus recommend the processing mechanism with the highest probability.
8. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.
9. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.
10. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the instruction is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
Citation Information
Patent Citations
Automatic classification method for social governance events based on neural network
CN113420015A
Basic-level governance event type identification method and system
CN113849595A
Intelligent event marking method and device and storage medium
CN116992034A
Semantic event detection using cross-domain knowledge
US20090299999A1
