Information comprehensive publishing method and system based on multi-source data fusion

Abstract

The present invention provides a method and system for comprehensive information release based on multi-source data fusion, which relates to the field of information processing technology. The method includes: obtaining the current feedback event sent by any grassroots platform, determining the target functional department; determining the current feedback state when determining that there is a first similar feedback event in the event processing library; when the current feedback state is a feedback state, retrieving the target feedback result and sending it to the grassroots platform; when the current feedback state is a non-feedback state, inputting the current feedback event into the multi-source data fusion set, continuously receiving the second similar feedback event, until receiving the information feedback result, and sending it to the grassroots platform corresponding to each event, so as to complete the comprehensive information release. For events that have not been fed back, the present invention will continuously accumulate the same or similar events and merge them for processing before giving the processing results of related events, further improving the work efficiency of department personnel and avoiding repeated and ineffective work.

Description

Technical Field

[0001] The present invention relates to the field of information processing technology, and in particular to a method and system for comprehensive information publishing based on multi-source data fusion. Background Art

[0002] In the current environment, there are many event processing systems that have been transferred from the urban area to the streets. Currently, not only at the street level, but also at the community and grid and other multi-source data platforms, there are various business systems. The data of the systems have not been shared, and no street-level thematic data set has been formed. In this case, when one or more people report the same event from multiple event platforms, the same event may be assigned to different departments for processing, or it may be continuously assigned to the same department for processing, depending on the different objects reflected in the event and the different objective descriptions of the event. After handling the event, each department needs to report materials with different templates for event feedback, which leads to heavy workload and heavy tasks for department personnel.

[0003] At the same time, the relevant platforms do not review duplicate events for reported events, and staff members need to spend a lot of time to verify, confirm and process these duplicate invalid events. In addition, each invalid event needs to report the processing results to superiors, which greatly reduces the work efficiency of staff members. For events involving multiple departments reported by the public, the feedback results of a single department are usually unable to solve the event problem. How to determine the department that needs to provide feedback on the event and how to enable the public to obtain feedback results from multiple departments in one feedback information are also important problems that plague the current situation.

[0004] However, there is currently no technical solution that can solve the above technical problems, and there is no comprehensive information publishing method and system based on multi-source data fusion. Summary of the invention

[0005] The present invention provides a method and system for comprehensive information release based on multi-source data fusion, which integrates multi-source platform data, comprehensively processes and releases current feedback events and related events, so as to improve the work efficiency of functional department staff.

[0006] In a first aspect, the present invention provides a method for comprehensive information release based on multi-source data fusion, comprising:

[0007] Obtaining a current feedback event sent by any grassroots platform, and determining a target functional department for handling the current feedback event according to the current feedback event, wherein the grassroots platform includes web platforms and software platforms of different levels with event feedback functions, and the different levels include street level, community level and grid level;

[0008] In the case where it is determined that there is a first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department, determining a current feedback state of the first similar feedback event;

[0009] In the case where the current feedback state is the feedback-already state, retrieving the target feedback result corresponding to the first similar feedback event from the event processing library, and sending the target feedback result as the feedback result of the current feedback event to the grassroots platform;

[0010] When the current feedback status is a non-feedback status, a multi-source data fusion set is established, the current feedback event is input into the multi-source data fusion set, and one or more second similar feedback events matching the current feedback event are continuously received until the information feedback result from the target functional department is received, and the information feedback result is sent to the grassroots platform corresponding to each event in the multi-source data fusion set to complete the comprehensive information release.

[0011] According to the information comprehensive publishing method based on multi-source data fusion provided by the present invention, before determining the current feedback status of the first similar feedback event, the method further includes:

[0012] Preprocessing the current feedback event and each candidate event in the event processing library to obtain all vector matching pairs, each vector matching pair including the current feedback event vector and any candidate event vector;

[0013] For any vector matching pair, a first condition matching value is determined according to the reserved contact information corresponding to the current feedback event and the reserved contact information corresponding to the candidate event; a second condition matching value is determined according to the reserved nickname corresponding to the current feedback event and the reserved nickname corresponding to the candidate event; a final similarity value is obtained according to the initial similarity value corresponding to the vector matching pair, the first condition matching value and the second condition matching value;

[0014] All vector matching pairs are traversed to obtain a final similarity value corresponding to each vector matching pair, and a candidate event corresponding to a vector matching pair whose final similarity value is greater than a first preset similarity is determined as a first similar feedback event matching the current feedback event.

[0015] According to the information comprehensive publishing method based on multi-source data fusion provided by the present invention, the obtaining of the final similarity value according to the initial similarity value corresponding to the vector matching pair, the first condition matching value and the second condition matching value includes:

[0016]

[0017] Among them, S is the final similarity value, is the initial similarity value, is the current feedback event vector, is any candidate event vector, is the modulus of the current feedback event vector, is the modulus of any candidate event vector, is the preset weight of the first condition matching value, is the preset weight of the second condition matching value, is the first condition matching value, is the second condition matching value;

[0018] Among them, when the reserved contact information corresponding to the current feedback event is the same as the reserved contact information corresponding to the candidate event, the first condition matching value is 1; when the reserved contact information corresponding to the current feedback event is different from the reserved contact information corresponding to the candidate event, the first condition matching value is 0; when the reserved nickname corresponding to the current feedback event is the same as the reserved nickname corresponding to the candidate event, the second condition matching value is 1; when the reserved nickname corresponding to the current feedback event is different from the reserved nickname corresponding to the candidate event, the second condition matching value is 0.

[0019] According to the information comprehensive publishing method based on multi-source data fusion provided by the present invention, after obtaining the final similarity value corresponding to each vector matching pair, the method further includes:

[0020] For any functional department, according to the current feedback event and each candidate event in the event processing library corresponding to the functional department, determine the event similarity value between the current feedback event and each candidate event;

[0021] Determine the candidate event corresponding to the event similarity value greater than the second preset similarity as the cross-department candidate event, wherein the second preset similarity is greater than or equal to the first preset similarity;

[0022] Traverse all functional departments, determine all cross-departmental events to be selected, determine the target functional department and the functional department corresponding to each cross-departmental event to be selected as the linkage department, and build a linkage information release window. The linkage information release window is used to receive feedback messages from each linkage department related to the current feedback event, and after obtaining the linkage feedback messages from all linkage departments, send the linkage feedback messages as the feedback results of the current feedback event to the grassroots platform to complete the comprehensive information release.

[0023] According to the information comprehensive publishing method based on multi-source data fusion provided by the present invention, the determining of the target functional department for handling the current feedback event according to the current feedback event includes:

[0024] Processing the input dialog box corresponding to the current feedback event to determine the target functional department;

[0025] In the case where the target functional department cannot be determined according to the input dialog box corresponding to the current feedback event, the current feedback event is processed according to preset keyword matching to determine the target functional department.

[0026] According to the comprehensive information release method based on multi-source data fusion provided by the present invention, the target functional departments at least include social security departments, urban and rural construction departments, environmental protection departments and health departments;

[0027] When the target functional department is a social security department, the preset keywords at least include employment, social security, wages, labor relations and talent introduction;

[0028] When the target functional department is the urban and rural construction department, the preset keywords at least include housing, real estate, property, urban planning, and building safety;

[0029] When the target functional department is an environmental protection department, the preset keywords at least include environmental protection, pollution, air quality, noise, and garbage classification;

[0030] When the target functional department is the health department, the preset keywords include at least medical care, health, vaccine, public health, and disease prevention.

[0031] According to the information comprehensive publishing method based on multi-source data fusion provided by the present invention, after processing the current feedback event according to preset keyword matching, the method further includes:

[0032] In the case where the target functional department cannot be determined according to the preset keyword matching, the current feedback event is input into the functional department classification model to obtain the target functional department corresponding to the current feedback event output by the functional department classification model;

[0033] The functional department classification model is determined after training based on all historical event corpus samples and the functional department label samples corresponding to each historical event corpus sample.

[0034] According to the information comprehensive publishing method based on multi-source data fusion provided by the present invention, before or after continuously receiving one or more second similar feedback events matching the current feedback event, the method further includes:

[0035] When the accumulated time duration of the unprocessed current feedback event is less than the preset time duration, and the number of events in the multi-source data fusion set is less than the preset number, marking the multi-source data fusion set as green;

[0036] When the cumulative time of the unprocessed current feedback event is greater than or equal to the preset time, and the number of events in the multi-source data fusion set is less than the preset number, or when the cumulative time of the unprocessed current feedback event is less than the preset time, and the number of events in the multi-source data fusion set is greater than or equal to the preset number, the multi-source data fusion set is marked as yellow, and an early warning instruction is generated, the early warning instruction is used to instruct the target functional department to process the multi-source data fusion set;

[0037] When the cumulative duration of the unprocessed current feedback event is greater than or equal to the preset duration, and the number of events in the multi-source data fusion set is greater than or equal to the preset number, the multi-source data fusion set is marked in red and an alarm instruction is generated. The alarm instruction is used to instruct the target functional department to process the multi-source data fusion set and send the multi-source data fusion set to the superior department of the target functional department.

[0038] According to the information comprehensive publishing method based on multi-source data fusion provided by the present invention, after determining the target functional department for handling the current feedback event according to the current feedback event, the method further includes:

[0039] When it is determined that there is no first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department, the current feedback event is stored in the event processing library, and an event instruction to be processed is generated;

[0040] The pending event instruction is used to instruct the target functional department to process the current feedback event.

[0041] In the second aspect, a comprehensive information publishing system based on multi-source data fusion is provided, comprising:

[0042] An acquisition unit, the acquisition unit is used to acquire a current feedback event sent by any grassroots platform, and determine a target functional department for processing the current feedback event according to the current feedback event, wherein the grassroots platform includes webpage platforms and software platforms of different levels with event feedback functions, and the different levels include street level, community level and grid level;

[0043] a determination unit, the determination unit being configured to determine a current feedback state of a first similar feedback event when it is determined that there is a first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department;

[0044] a first sending unit, configured to retrieve, when the current feedback state is a feedback-accepted state, a target feedback result corresponding to the first similar feedback event from the event processing library, and send the target feedback result as a feedback result of the current feedback event to the grassroots platform;

[0045] The second sending unit is used to establish a multi-source data fusion set when the current feedback state is a non-feedback state, input the current feedback event into the multi-source data fusion set, and continuously receive one or more second similar feedback events that match the current feedback event until receiving the information feedback result from the target functional department, and send the information feedback result to the grassroots platform corresponding to each event in the multi-source data fusion set to complete the comprehensive information release.

[0046] The present invention can timely receive event feedback from grassroots platforms at different levels (street level, community level, grid level), provide multi-source data for subsequent event processing and analysis, and automatically assign events to the most appropriate functional departments for processing through preset judgment steps and intelligent analysis. In the processing process, different processing strategies will be executed according to the current feedback status of the first similar feedback event that matches the current feedback event. For events that have been fed back, the feedback of the current feedback event is directly realized according to the processing result of the first similar feedback event, so there is no need to repeatedly investigate, analyze, and answer the same event questions, thereby improving the work efficiency of department personnel; for events that have not been fed back, before giving the processing results of related events, the same or similar events will be continuously accumulated and merged for processing to realize comprehensive information release, further improving the work efficiency of department personnel and avoiding repeated ineffective work;

[0047] In the process of identifying similar feedback events, the present invention fully combines the cosine similarity algorithm and fully considers the influence of the reserved contact information and the reserved nickname on the identification of similar feedback events, so as to accurately and effectively determine the first similar feedback event that matches the current feedback event;

[0048] The present invention can also conduct cross-department linkage event feedback, determine the target functional department and the functional department corresponding to each cross-departmental event to be selected as the linkage department, and use the linkage information release window to realize the event feedback of the linkage department, thereby improving the credibility of the event feedback information and enhancing the public's trust and satisfaction;

[0049] The present invention also determines the target functional department for handling the current feedback event through preset keyword matching with a sequential execution order and a functional department classification model, which not only improves the accuracy of determining the target functional department, but also improves the robustness of determining the target functional department through different processing strategies; the present invention also combines the two processing dimensions of the cumulative duration of the current feedback event and the number of events in the multi-source data fusion set, and executes different response strategies in different situations, so as to help the functional department personnel to intuitively understand the processing progress and severity of the current feedback event, timely plan the current work arrangements, and reduce the risk of work overdue. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0051] Figure 1 It is a flow chart of the comprehensive information publishing method based on multi-source data fusion provided by the present invention;

[0052] Figure 2 It is a structural schematic diagram of the information comprehensive publishing system based on multi-source data fusion provided by the present invention;

[0053] Figure 3 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Figure 1 : is a flow chart of a method for comprehensive information release based on multi-source data fusion provided by the present invention, wherein the method for comprehensive information release based on multi-source data fusion comprises:

[0056] Step 101: obtaining a current feedback event sent by any grassroots platform, and determining a target functional department for processing the current feedback event according to the current feedback event, wherein the grassroots platform includes web platforms and software platforms of different levels with event feedback functions, and the different levels include street level, community level and grid level;

[0057] Step 102: when it is determined that there is a first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department, determine the current feedback status of the first similar feedback event;

[0058] Step 103: when the current feedback state is the feedback-already state, retrieve the target feedback result corresponding to the first similar feedback event from the event processing library, and send the target feedback result as the feedback result of the current feedback event to the grassroots platform;

[0059] Step 104: When the current feedback state is a non-feedback state, a multi-source data fusion set is established, the current feedback event is input into the multi-source data fusion set, and one or more second similar feedback events matching the current feedback event are continuously received until the information feedback result from the target functional department is received, and the information feedback result is sent to the grassroots platform corresponding to each event in the multi-source data fusion set to complete the comprehensive information release.

[0060] In step 101, the present invention first receives current feedback events sent from grassroots platforms at different levels (including street level, community level and grid level). These grassroots platforms usually have event feedback functions, which can be online or offline feedback modes, allowing grassroots people to submit problems or suggestions encountered. By analyzing the content of these feedback events, using a preset keyword library or natural language processing technology, the corresponding functional departments can be identified, and the processing department of the feedback event can be automatically determined by technical means, reducing the workload of manual judgment and allocation, improving processing efficiency, ensuring that each feedback event can be correctly and quickly allocated to the corresponding functional department, and avoiding processing delays or invalid processing caused by allocation errors; at the same time, the present invention fully sinks the event feedback within the jurisdiction of the functional department to different levels, breaking through the independence of existing feedback between different levels, and realizing information aggregation of multi-source data.

[0061] In step 102, after determining the target functional department, the present invention will query the event processing library corresponding to the functional department to find the first similar feedback event that matches the current feedback event. These similar events may be cases that have been processed before and recorded in the library. The present invention can use the file processing method of cosine similarity, combined with event similarities under different dimensions such as reserved contact information, reserved nicknames, geographical locations for publishing information, and time intervals for publishing information, to comprehensively determine whether there is a first similar feedback event that matches the current feedback event in the event processing library corresponding to the target functional department. When it is determined that there is a first similar feedback event that matches the current feedback event in the event processing library corresponding to the target functional department, the current feedback status of the first similar feedback event is further determined, and the current feedback status includes whether it has been processed or is still in a pending state.

[0062] Optionally, after determining a target functional department for handling the current feedback event according to the current feedback event, the method further includes:

[0063] When it is determined that there is no first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department, the current feedback event is stored in the event processing library, and an event instruction to be processed is generated;

[0064] The pending event instruction is used to instruct the target functional department to process the current feedback event.

[0065] Optionally, the current feedback event is stored in the event processing library, and after obtaining relevant feedback results, it provides a reference for the subsequent processing of similar events, which helps to improve the accuracy and consistency of event processing.

[0066] In step 103, if the current feedback status of the first similar feedback event is the feedback status, that is, the event has been processed before and has a clear result, the present invention will retrieve the feedback result corresponding to the first similar feedback event from the event processing library, and send this feedback result as the feedback result of the current feedback event to the grassroots platform. In this case, since the direct intervention of the functional department personnel can be reduced, not only the processing efficiency of the current feedback event is improved, but also the public can be provided with instant processing result feedback, thereby improving user experience and satisfaction.

[0067] In step 104, if the current feedback status of the first similar feedback event is a non-feedback status, a multi-source data fusion set will be established based on the first similar feedback event, and the current feedback event will be input into the set. Before the functional department personnel have time to process these feedback events, the present invention will continue to receive one or more second similar feedback events that match the current feedback event. These similar feedback events may come from different grassroots platforms or subsequent feedback on the same platform. They may be feedback on the same topic or different topics by the same person, or feedback on the same topic or different topics by different people, until the information feedback results from the target functional department personnel are received, and the information feedback results are immediately sent to the grassroots platform corresponding to each event in the multi-source data fusion set to complete the comprehensive information release.

[0068] Optionally, staff members of the target functional departments can also achieve a comprehensive understanding and comprehensive analysis of similar incidents through multi-source data fusion, provide stronger support for decision-making, continuously receive and process relevant feedback events, and ensure that problems are fully and thoroughly resolved.

[0069] The present invention can timely receive event feedback from grassroots platforms at different levels (street level, community level, grid level), provide multi-source data for subsequent event processing and analysis, and automatically assign events to the most appropriate functional department for processing through preset judgment steps and intelligent analysis. During the processing, different processing strategies will be executed according to the current feedback status of the first similar feedback event that matches the current feedback event. For events that have been fed back, the feedback of the current feedback event is directly realized according to the processing result of the first similar feedback event, thereby eliminating the need for repeated investigation, analysis, and answering of the same event questions, thereby improving the work efficiency of department personnel; for events that have not been fed back, the same or similar events will be continuously accumulated and merged for processing before the processing results of related events are given, so as to realize comprehensive information release, further improve the work efficiency of department personnel, and avoid repeated and ineffective work.

[0070] Optionally, before determining the current feedback state of the first similar feedback event, the method further includes:

[0071] Preprocessing the current feedback event and each candidate event in the event processing library to obtain all vector matching pairs, each vector matching pair including the current feedback event vector and any candidate event vector;

[0072] For any vector matching pair, a first condition matching value is determined according to the reserved contact information corresponding to the current feedback event and the reserved contact information corresponding to the candidate event; a second condition matching value is determined according to the reserved nickname corresponding to the current feedback event and the reserved nickname corresponding to the candidate event; a final similarity value is obtained according to the initial similarity value corresponding to the vector matching pair, the first condition matching value and the second condition matching value;

[0073] All vector matching pairs are traversed to obtain a final similarity value corresponding to each vector matching pair, and a candidate event corresponding to a vector matching pair whose final similarity value is greater than a first preset similarity is determined as a first similar feedback event matching the current feedback event.

[0074] Optionally, the present invention cleans the description text of the current feedback event and each candidate event in the event processing library, removes irrelevant symbols, stop words and non-text content, and uses a preset natural language processing technology to convert the cleaned text into a vector in a high-dimensional vector space, namely, the current feedback event vector and the candidate event vector. For each candidate event in the event processing library, a vector matching pair is formed with the current feedback event to facilitate subsequent calculation of similarity.

[0075] Furthermore, the reserved contact information in the current feedback event and the candidate event, such as telephone number, email address, etc., is extracted. If the feedback contains the user's nickname or name, the nickname is further matched and the second condition matching value is calculated. Optionally, when the reserved contact information corresponding to the current feedback event is the same as the reserved contact information corresponding to the candidate event, the first condition matching value is 1; when the reserved contact information corresponding to the current feedback event is different from the reserved contact information corresponding to the candidate event, the first condition matching value is 0; when the reserved nickname corresponding to the current feedback event is the same as the reserved nickname corresponding to the candidate event, the second condition matching value is 1; when the reserved nickname corresponding to the current feedback event is different from the reserved nickname corresponding to the candidate event, the second condition matching value is 0.

[0076] Furthermore, obtaining a final similarity value according to the initial similarity value corresponding to the vector matching pair, the first condition matching value, and the second condition matching value includes:

[0077] (1)

[0078] Among them, S is the final similarity value, is the initial similarity value, is the current feedback event vector, is any candidate event vector, is the modulus of the current feedback event vector, is the modulus of any candidate event vector, is the preset weight of the first condition matching value, is the preset weight of the second condition matching value, is the first condition matching value, is the second condition matching value. Through the introduction of vectorized representation and condition matching value, not only the similarity of text content is considered, but also additional information such as user contact information and nicknames are integrated, making the matching of similar events more comprehensive and accurate.

[0079] Optionally, based on the vectorized text content, the initial similarity value is calculated using cosine similarity, and then the initial similarity value, the first condition matching value, and the second condition matching value are weightedly fused to obtain the final similarity value, and all vector matching pairs are traversed to obtain the final similarity value corresponding to each vector matching pair. Optionally, the first preset similarity can be 0.8, and the candidate event corresponding to the vector matching pair with the final similarity value greater than 0.8 is determined as the first similar feedback event that matches the current feedback event. In the process of identifying similar feedback events, the present invention fully combines the cosine similarity algorithm, and fully considers the influence of reserved contact information and reserved nicknames on the identification of similar feedback events, and accurately and effectively determines the first similar feedback event that matches the current feedback event.

[0080] Optionally, after obtaining the final similarity value corresponding to each vector matching pair, the method further includes:

[0081] For any functional department, according to the current feedback event and each candidate event in the event processing library corresponding to the functional department, determine the event similarity value between the current feedback event and each candidate event;

[0082] Determine the candidate event corresponding to the event similarity value greater than the second preset similarity as the cross-department candidate event, wherein the second preset similarity is greater than or equal to the first preset similarity;

[0083] Traverse all functional departments, determine all cross-departmental events to be selected, determine the target functional department and the functional department corresponding to each cross-departmental event to be selected as the linkage department, and build a linkage information release window. The linkage information release window is used to receive feedback messages from each linkage department related to the current feedback event, and after obtaining the linkage feedback messages from all linkage departments, send the linkage feedback messages as the feedback results of the current feedback event to the grassroots platform to complete the comprehensive information release.

[0084] Optionally, since the current feedback event may involve feedback from multiple departments, the feedback result of a single department cannot completely solve the problems of the masses. At this time, it is necessary to confirm the associated department corresponding to the current feedback event. The present invention can call the event processing library in all functional departments, and for each event processing library, pre-process the current feedback event and each candidate event in the event processing library to obtain all vector matching pairs, each vector matching pair includes the current feedback event vector and any candidate event vector; for any vector matching pair, determine the first condition matching value according to the reserved contact method corresponding to the current feedback event and the reserved contact method corresponding to the candidate event; Determine the second condition matching value according to the reserved nickname corresponding to the current feedback event and the reserved nickname corresponding to the candidate event; obtain the event similarity value according to the initial similarity value corresponding to the vector matching pair, the first condition matching value and the second condition matching value; traverse each vector matching pair in all event processing libraries to obtain the event similarity value corresponding to each vector matching pair, and determine the candidate event corresponding to the event similarity value greater than the second preset similarity as a cross-department candidate event, wherein the second preset similarity is greater than or equal to the first preset similarity. For example, if the first preset similarity is set to 0.75, the second preset similarity is set to 0.95.

[0085] Further, the second preset similarity is used to screen all event similarity values, determine all cross-departmental selected events, determine the target functional department and the functional department corresponding to each cross-departmental selected event as a linkage department, and construct a linkage information release window, the linkage information release window is used to receive the feedback message of each linkage department related to the current feedback event, automatically integrate these feedbacks, and form a comprehensive feedback report, the report can include the processing opinions, processing results, existing problems and improvement suggestions of each department, and after obtaining the linkage feedback message of all linkage departments, the linkage feedback message is sent to the grassroots platform as the feedback result of the current feedback event to complete the comprehensive information release. The present invention conducts linkage event feedback across departments, determines the target functional department and the functional department corresponding to each cross-departmental selected event as a linkage department, can monitor the task processing progress of each linkage department, automatically sends a reminder notification to the department with lagging progress, urges it to speed up the processing speed, obtains a comprehensive feedback report by constructing a linkage information release window, and uses the linkage information release window to realize the event feedback of the linkage department, ensuring that each current feedback event can be properly handled, thereby improving the credibility of event feedback information, and improving the trust and satisfaction of the masses.

[0086] Optionally, determining a target functional department for handling the current feedback event according to the current feedback event includes:

[0087] Processing the input dialog box corresponding to the current feedback event to determine the target functional department;

[0088] In the case where the target functional department cannot be determined according to the input dialog box corresponding to the current feedback event, the current feedback event is processed according to preset keyword matching to determine the target functional department.

[0089] Optionally, in addition to the basic text input box, a drop-down selection box or label selection function is added to the input dialog box for receiving public feedback, allowing the public to directly select or mark the functional department related to the problem when submitting feedback. For example, set the "problem type" or "related department" option to list common problem categories and corresponding functional departments, such as "human resources and social security", "housing and urban and rural construction", "environmental protection", "health and health", etc., and then directly determine the target functional department; in the case that the target functional department cannot be determined in the input dialog box corresponding to the current feedback event, the current feedback event is processed according to the preset keyword matching to determine the target functional department, and the processed text is matched with the preset keyword library. String matching algorithms (such as fuzzy matching, regular expression matching, etc.) can be used to achieve this. The target functional department is determined according to the keyword with the highest matching degree. If there are multiple high-matching keywords, the final functional department is determined according to the preset priority rules or further analysis of the text content.

[0090] Optionally, the target functional departments include at least social security departments, urban and rural construction departments, environmental protection departments and health departments;

[0091] When the target functional department is a social security department, the preset keywords at least include employment, social security, wages, labor relations and talent introduction;

[0092] When the target functional department is the urban and rural construction department, the preset keywords at least include housing, real estate, property, urban planning, and building safety;

[0093] When the target functional department is an environmental protection department, the preset keywords at least include environmental protection, pollution, air quality, noise, and garbage classification;

[0094] When the target functional department is the health department, the preset keywords include at least medical care, health, vaccine, public health, and disease prevention. In other embodiments, other functional departments may also be included, which will not be elaborated here.

[0095] Optionally, after processing the current feedback event according to preset keyword matching, the method further includes:

[0096] In the case where the target functional department cannot be determined according to the preset keyword matching, the current feedback event is input into the functional department classification model to obtain the target functional department corresponding to the current feedback event output by the functional department classification model;

[0097] The functional department classification model is determined after training based on all historical event corpus samples and the functional department label samples corresponding to each historical event corpus sample.

[0098] Optionally, the present invention, based on the use of preset keyword matching, introduces a functional department classification model as a supplementary mechanism in order to further improve the accuracy and flexibility of feedback event processing. When the target functional department cannot be determined directly through the preset keyword library matching, the functional department classification model is used to automatically intervene and conduct in-depth analysis and prediction of the current feedback event that cannot be directly classified. The present invention uses the collected historical event corpus samples and their functional department label samples, and adopts machine learning or deep learning algorithms (such as text classification algorithms, neural networks, etc.) to train the functional department classification model. Through continuous iterative optimization, the classification accuracy and generalization ability of the model are improved, so that after the current feedback event is input into the functional department classification model, the target functional department corresponding to the current feedback event output by the functional department classification model can be obtained.

[0099] The present invention introduces a functional department classification model, which can quickly and accurately determine the target functional department when it is impossible to directly match the preset keywords, reduce manual intervention and waiting time, and improve the processing efficiency of feedback events; through preset keyword matching with a sequential execution order and a functional department classification model, the target functional department for processing the current feedback event is determined, thereby improving the robustness of the determination of the target functional department.

[0100] Optionally, before or after continuously receiving one or more second similar feedback events matching the current feedback event, the method further includes:

[0101] When the accumulated time duration of the unprocessed current feedback event is less than the preset time duration, and the number of events in the multi-source data fusion set is less than the preset number, marking the multi-source data fusion set as green;

[0102] When the cumulative time of the unprocessed current feedback event is greater than or equal to the preset time, and the number of events in the multi-source data fusion set is less than the preset number, or when the cumulative time of the unprocessed current feedback event is less than the preset time, and the number of events in the multi-source data fusion set is greater than or equal to the preset number, the multi-source data fusion set is marked as yellow, and an early warning instruction is generated, the early warning instruction is used to instruct the target functional department to process the multi-source data fusion set;

[0103] When the cumulative duration of the unprocessed current feedback event is greater than or equal to the preset duration, and the number of events in the multi-source data fusion set is greater than or equal to the preset number, the multi-source data fusion set is marked in red and an alarm instruction is generated. The alarm instruction is used to instruct the target functional department to process the multi-source data fusion set and send the multi-source data fusion set to the superior department of the target functional department.

[0104] Optionally, the present invention continuously tracks the processing status of the multi-source data fusion set by establishing a real-time monitoring and feedback mechanism. When the processing status changes, the system promptly updates the status indicator (green, yellow, red), and sends update information to relevant functional departments and their superior departments through various channels (such as system notifications, text messages, emails, etc.) to ensure the transparency and timeliness of the information. It also effectively shortens the processing time of feedback events and improves processing efficiency through dynamic threshold adjustment and event priority evaluation. By combining the two processing dimensions of the cumulative duration of the current feedback event and the number of events in the multi-source data fusion set, different response strategies are executed in different situations to help functional department personnel intuitively understand the processing progress and severity of the current feedback event, plan current work arrangements in a timely manner, and reduce the risk of overdue work.

[0105] Figure 2 It is a structural diagram of the information comprehensive publishing system based on multi-source data fusion provided by the present invention, the information comprehensive publishing system based on multi-source data fusion includes an acquisition unit 1, the acquisition unit is used to acquire the current feedback event sent by any grassroots platform, and determine the target functional department for processing the current feedback event according to the current feedback event. The grassroots platform includes web platforms and software platforms of different levels with event feedback functions, and the different levels include street level, community level and grid level. The working principle of the acquisition unit 1 can refer to the aforementioned step 101 and will not be repeated here.

[0106] The comprehensive information publishing system based on multi-source data fusion also includes a determination unit 2, which is used to determine the current feedback status of the first similar feedback event when it is determined that there is a first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department. The working principle of the determination unit 2 can refer to the aforementioned step 102 and will not be repeated here.

[0107] The information comprehensive publishing system based on multi-source data fusion also includes a first sending unit 3, which is used to retrieve the target feedback result corresponding to the first similar feedback event from the event processing library when the current feedback state is the feedback-accepted state, and send the target feedback result as the feedback result of the current feedback event to the grassroots platform. The working principle of the first sending unit 3 can refer to the aforementioned step 103 and will not be repeated here.

[0108] The comprehensive information release system based on multi-source data fusion also includes a second sending unit 4, which is used to establish a multi-source data fusion set when the current feedback state is a non-feedback state, input the current feedback event into the multi-source data fusion set, and continuously receive one or more second similar feedback events matching the current feedback event until receiving the information feedback result from the target functional department, and send the information feedback result to the grassroots platform corresponding to each event in the multi-source data fusion set to complete the comprehensive information release. The working principle of the second sending unit 4 can refer to the aforementioned step 104 and will not be repeated here.

[0109] The present invention can timely receive event feedback from grassroots platforms at different levels (street level, community level, grid level), provide multi-source data for subsequent event processing and analysis, and automatically assign events to the most appropriate functional department for processing through preset judgment steps and intelligent analysis. During the processing, different processing strategies will be executed according to the current feedback status of the first similar feedback event that matches the current feedback event. For events that have been fed back, the feedback of the current feedback event is directly realized according to the processing result of the first similar feedback event, thereby eliminating the need for repeated investigation, analysis, and answering of the same event questions, thereby improving the work efficiency of department personnel; for events that have not been fed back, the same or similar events will be continuously accumulated and merged for processing before the processing results of related events are given, so as to realize comprehensive information release, further improve the work efficiency of department personnel, and avoid repeated and ineffective work.

[0110] Figure 3 Schematic diagram of the structure of the electronic device provided by the present invention. Figure 3As shown, the electronic device may include: a processor 310, a communications interface 320, a memory 330 and a communications bus 340, wherein the processor 310, the communications interface 320 and the memory 330 communicate with each other through the communications bus 340. The processor 310 may call the logic instructions in the memory 330 to execute the information comprehensive release method based on multi-source data fusion, the method comprising: obtaining the current feedback event sent by any grassroots platform, and determining the target functional department for processing the current feedback event according to the current feedback event, wherein the grassroots platform comprises webpage platforms and software platforms of different levels with event feedback functions, wherein the different levels include street level, community level and grid level; in the case of determining that there is a first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department, determining the current feedback state of the first similar feedback event; and when the current feedback state is In the case of the feedback received state, the target feedback result corresponding to the first similar feedback event is retrieved from the event processing library, and the target feedback result is sent to the grassroots platform as the feedback result of the current feedback event; when the current feedback state is the non-feedback state, a multi-source data fusion set is established, the current feedback event is input into the multi-source data fusion set, and one or more second similar feedback events matching the current feedback event are continuously received until the information feedback result from the target functional department is received, and the information feedback result is sent to the grassroots platform corresponding to each event in the multi-source data fusion set to complete the comprehensive information release.

[0111] In addition, the logic instructions in the above-mentioned memory 330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0112] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute a comprehensive information publishing method based on multi-source data fusion provided by the above methods, the method including: obtaining a current feedback event sent by any grassroots platform, and determining a target functional department for processing the current feedback event based on the current feedback event, the grassroots platform including web platforms and software platforms of different levels with event feedback functions, the different levels including street level, community level and grid level; determining whether there is a first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department In this case, determine the current feedback status of the first similar feedback event; when the current feedback status is the feedback-received status, retrieve the target feedback result corresponding to the first similar feedback event from the event processing library, and send the target feedback result as the feedback result of the current feedback event to the grassroots platform; when the current feedback status is the non-feedback status, establish a multi-source data fusion set, input the current feedback event into the multi-source data fusion set, continuously receive one or more second similar feedback events matching the current feedback event, until the information feedback result from the target functional department is received, and send the information feedback result to the grassroots platform corresponding to each event in the multi-source data fusion set to complete the comprehensive information release.

[0113] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the information comprehensive publishing method based on multi-source data fusion provided by the above-mentioned methods, the method comprising: obtaining a current feedback event sent by any grassroots platform, and determining a target functional department for processing the current feedback event according to the current feedback event, wherein the grassroots platform comprises webpage platforms and software platforms of different levels with event feedback functions, and the different levels comprise street level, community level and grid level; in the case of determining that a first similar feedback event matching the current feedback event exists in the event processing library corresponding to the target functional department, determining the first similar feedback event; the current feedback status of the event; when the current feedback status is the feedback-received status, the target feedback result corresponding to the first similar feedback event is retrieved from the event processing library, and the target feedback result is sent to the grassroots platform as the feedback result of the current feedback event; when the current feedback status is the non-feedback status, a multi-source data fusion set is established, the current feedback event is input into the multi-source data fusion set, and one or more second similar feedback events matching the current feedback event are continuously received until the information feedback result from the target functional department is received, and the information feedback result is sent to the grassroots platform corresponding to each event in the multi-source data fusion set to complete the comprehensive information release.

[0114] The device embodiments described above are merely illustrative, 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 scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0115] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0116] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A comprehensive information publishing method based on multi-source data fusion, characterized in that: include: Obtaining a current feedback event sent by any grassroots platform, and determining a target functional department for handling the current feedback event according to the current feedback event, wherein the grassroots platform includes web platforms and software platforms of different levels with event feedback functions, and the different levels include street level, community level and grid level; In the case where it is determined that there is a first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department, determining a current feedback state of the first similar feedback event; In the case where the current feedback state is the feedback-already state, retrieving the target feedback result corresponding to the first similar feedback event from the event processing library, and sending the target feedback result as the feedback result of the current feedback event to the grassroots platform; In the case where the current feedback state is a non-feedback state, a multi-source data fusion set is established, the current feedback event is input into the multi-source data fusion set, one or more second similar feedback events matching the current feedback event are continuously received, until the information feedback result from the target functional department is received, and the information feedback result is sent to the grassroots platform corresponding to each event in the multi-source data fusion set to complete the comprehensive information release; Before determining the current feedback state of the first similar feedback event, the method further includes: Preprocessing the current feedback event and each candidate event in the event processing library to obtain all vector matching pairs, each vector matching pair including the current feedback event vector and any candidate event vector; For any vector matching pair, a first condition matching value is determined according to the reserved contact information corresponding to the current feedback event and the reserved contact information corresponding to the candidate event; a second condition matching value is determined according to the reserved nickname corresponding to the current feedback event and the reserved nickname corresponding to the candidate event; a final similarity value is obtained according to the initial similarity value corresponding to the vector matching pair, the first condition matching value and the second condition matching value; Traversing all vector matching pairs to obtain a final similarity value corresponding to each vector matching pair, and determining a candidate event corresponding to a vector matching pair whose final similarity value is greater than a first preset similarity as a first similar feedback event matching the current feedback event; After obtaining the final similarity value corresponding to each vector matching pair, the method further includes: For any functional department, according to the current feedback event and each candidate event in the event processing library corresponding to the functional department, determine the event similarity value between the current feedback event and each candidate event; Determine the candidate event corresponding to the event similarity value greater than the second preset similarity as the cross-department candidate event, wherein the second preset similarity is greater than or equal to the first preset similarity; Traverse all functional departments, determine all cross-departmental events to be selected, determine the target functional department and the functional department corresponding to each cross-departmental event to be selected as the linkage department, and build a linkage information release window, the linkage information release window is used to receive feedback messages from each linkage department related to the current feedback event, and after obtaining the linkage feedback messages from all linkage departments, send the linkage feedback messages as the feedback results of the current feedback event to the grassroots platform to complete the comprehensive information release; Determining a target functional department for handling the current feedback event according to the current feedback event includes: Processing the input dialog box corresponding to the current feedback event to determine the target functional department; In the case where the target functional department cannot be determined according to the input dialog box corresponding to the current feedback event, the current feedback event is processed according to preset keyword matching to determine the target functional department; After processing the current feedback event according to the preset keyword matching, the method further includes: In the case where the target functional department cannot be determined according to the preset keyword matching, the current feedback event is input into the functional department classification model to obtain the target functional department corresponding to the current feedback event output by the functional department classification model; The functional department classification model is determined after training based on all historical event corpus samples and the functional department label samples corresponding to each historical event corpus sample.

2. The method for comprehensive information release based on multi-source data fusion according to claim 1 is characterized in that: The obtaining a final similarity value according to the initial similarity value corresponding to the vector matching pair, the first condition matching value, and the second condition matching value includes: ； Among them, S is the final similarity value, is the initial similarity value, is the current feedback event vector, is any candidate event vector, is the modulus of the current feedback event vector, is the modulus of any candidate event vector, is the preset weight of the first condition matching value, is the preset weight of the second condition matching value, is the first condition matching value, is the second condition matching value; Among them, when the reserved contact information corresponding to the current feedback event is the same as the reserved contact information corresponding to the candidate event, the first condition matching value is 1; when the reserved contact information corresponding to the current feedback event is different from the reserved contact information corresponding to the candidate event, the first condition matching value is 0; when the reserved nickname corresponding to the current feedback event is the same as the reserved nickname corresponding to the candidate event, the second condition matching value is 1; when the reserved nickname corresponding to the current feedback event is different from the reserved nickname corresponding to the candidate event, the second condition matching value is 0.

3. The method for comprehensive information release based on multi-source data fusion according to claim 1 is characterized in that: The target functional departments include at least the social security department, urban and rural construction department, environmental protection department and health department; When the target functional department is a social security department, the preset keywords at least include employment, social security, wages, labor relations and talent introduction; When the target functional department is the urban and rural construction department, the preset keywords at least include housing, real estate, property, urban planning, and building safety; When the target functional department is an environmental protection department, the preset keywords at least include environmental protection, pollution, air quality, noise, and garbage classification; When the target functional department is the health department, the preset keywords include at least medical care, health, vaccine, public health, and disease prevention.

4. The method for comprehensive information release based on multi-source data fusion according to claim 1 is characterized in that: Before or after continuously receiving one or more second similar feedback events matching the current feedback event, the method further includes: When the accumulated time duration of the unprocessed current feedback event is less than the preset time duration, and the number of events in the multi-source data fusion set is less than the preset number, marking the multi-source data fusion set as green; When the cumulative time of the unprocessed current feedback event is greater than or equal to the preset time, and the number of events in the multi-source data fusion set is less than the preset number, or when the cumulative time of the unprocessed current feedback event is less than the preset time, and the number of events in the multi-source data fusion set is greater than or equal to the preset number, the multi-source data fusion set is marked as yellow, and an early warning instruction is generated, the early warning instruction is used to instruct the target functional department to process the multi-source data fusion set; When the cumulative duration of the unprocessed current feedback event is greater than or equal to the preset duration, and the number of events in the multi-source data fusion set is greater than or equal to the preset number, the multi-source data fusion set is marked in red and an alarm instruction is generated. The alarm instruction is used to instruct the target functional department to process the multi-source data fusion set and send the multi-source data fusion set to the superior department of the target functional department.

5. The method for comprehensive information release based on multi-source data fusion according to claim 1 is characterized in that: After determining the target functional department for handling the current feedback event according to the current feedback event, the method further includes: When it is determined that there is no first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department, the current feedback event is stored in the event processing library, and an event instruction to be processed is generated; The pending event instruction is used to instruct the target functional department to process the current feedback event.

6. An information comprehensive publishing system based on multi-source data fusion, characterized in that: include: An acquisition unit, the acquisition unit is used to acquire a current feedback event sent by any grassroots platform, and determine a target functional department for processing the current feedback event according to the current feedback event, wherein the grassroots platform includes webpage platforms and software platforms of different levels with event feedback functions, and the different levels include street level, community level and grid level; a determination unit, the determination unit being configured to determine a current feedback state of a first similar feedback event when it is determined that there is a first similar feedback event matching the current feedback event in the event processing library corresponding to the target functional department; a first sending unit, configured to retrieve, when the current feedback state is a feedback-accepted state, a target feedback result corresponding to the first similar feedback event from the event processing library, and send the target feedback result as a feedback result of the current feedback event to the grassroots platform; A second sending unit, the second sending unit is used to establish a multi-source data fusion set when the current feedback state is a non-feedback state, input the current feedback event into the multi-source data fusion set, continuously receive one or more second similar feedback events matching the current feedback event, until receiving the information feedback result from the target functional department, and send the information feedback result to the grassroots platform corresponding to each event in the multi-source data fusion set to complete the comprehensive information release; Before determining the current feedback state of the first similar feedback event, the method further includes: Preprocessing the current feedback event and each candidate event in the event processing library to obtain all vector matching pairs, each vector matching pair including the current feedback event vector and any candidate event vector; For any vector matching pair, a first condition matching value is determined according to the reserved contact information corresponding to the current feedback event and the reserved contact information corresponding to the candidate event; a second condition matching value is determined according to the reserved nickname corresponding to the current feedback event and the reserved nickname corresponding to the candidate event; a final similarity value is obtained according to the initial similarity value corresponding to the vector matching pair, the first condition matching value and the second condition matching value; Traversing all vector matching pairs to obtain a final similarity value corresponding to each vector matching pair, and determining a candidate event corresponding to a vector matching pair whose final similarity value is greater than a first preset similarity as a first similar feedback event matching the current feedback event; After obtaining the final similarity value corresponding to each vector matching pair, it also includes: For any functional department, according to the current feedback event and each candidate event in the event processing library corresponding to the functional department, determine the event similarity value between the current feedback event and each candidate event; Determine the candidate event corresponding to the event similarity value greater than the second preset similarity as the cross-department candidate event, wherein the second preset similarity is greater than or equal to the first preset similarity; Traverse all functional departments, determine all cross-departmental events to be selected, determine the target functional department and the functional department corresponding to each cross-departmental event to be selected as the linkage department, and build a linkage information release window, the linkage information release window is used to receive feedback messages from each linkage department related to the current feedback event, and after obtaining the linkage feedback messages from all linkage departments, send the linkage feedback messages as the feedback results of the current feedback event to the grassroots platform to complete the comprehensive information release; Determining a target functional department for handling the current feedback event according to the current feedback event includes: Processing the input dialog box corresponding to the current feedback event to determine the target functional department; In the case where the target functional department cannot be determined according to the input dialog box corresponding to the current feedback event, the current feedback event is processed according to preset keyword matching to determine the target functional department; After processing the current feedback event according to the preset keyword matching, the method further includes: In the case where the target functional department cannot be determined according to the preset keyword matching, the current feedback event is input into the functional department classification model to obtain the target functional department corresponding to the current feedback event output by the functional department classification model; The functional department classification model is determined after training based on all historical event corpus samples and the functional department label samples corresponding to each historical event corpus sample.

Images