Base-level report fusion system and method

Abstract

The invention discloses a basic-level report fusion system and method, belongs to the technical field of data processing, and aims to solve the technical problem of how to optimize a data acquisition and fusion process and efficiently and accurately provide a data report. Comprising a report issuing module used for uploading data to a big data supervisor module in a report form; the big data supervisor module is used for summarizing the reports uploaded by the report issuing departments, performing information extraction, data analysis and standardization processing on the reports to obtain a standing book, and uploading the standing book to the report reporting module; the report reporting module is used for supporting a report reporting department to fill and report the standing book which passes the confirmation, and for the standing book which does not pass the confirmation, the report reporting module is used for returning a modification opinion fed back by the report reporting department to the big data supervisor module; correspondingly, the big data competent module is used for supporting the big data competent department to modify the machine account based on the modification opinions.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a grassroots report fusion system and method. Background Art

[0002] With the in-depth development of the information age, grassroots staff are facing unprecedented challenges. They need to deal with a large number of form filling tasks from many superior departments. This is not only a huge workload, but also seriously occupies the time that should be used for field research, thus affecting the efficiency improvement of grassroots work. At present, grassroots staff mainly face the following problems:

[0003] (1) The task of filling out forms is arduous and involves a wide range of business lines: Grassroots staff need to meet the data needs of dozens of superior departments and units at the same time. These departments frequently and in large quantities require data, which forces grassroots staff to devote a lot of energy to complete the tedious report filling work;

[0004] (2) Multiple entries and repeated reporting: Since the reporting systems are scattered across various business lines and there is a lack of effective coordination between different superior departments and units, the content and format of the reports issued are not unified, which in turn causes grassroots personnel to have to fill in repeated reports;

[0005] (3) Poor data sharing affects work efficiency: Data barriers and information islands between departments are serious. Grassroots staff need to manually integrate information from various departments when reporting data, which not only increases the workload but also easily leads to information omissions and deviations.

[0006] For form filling tasks, how to optimize the data collection and integration process and provide data reports efficiently and accurately is a technical problem that needs to be solved. Summary of the invention

[0007] The technical task of the present invention is to address the above deficiencies and provide a grassroots report fusion system and method to solve the technical problem of how to optimize data collection and fusion processes and provide data reports efficiently and accurately.

[0008] In a first aspect, the present invention provides a grassroots report fusion system, including a report issuing module, a big data management module and a report submitting module;

[0009] The report issuing module communicates with the report issuing department as a user through the report issuing interface, and is used to support each report issuing department to upload the daily collected data, and upload the data to the big data supervisor module in the form of reports;

[0010] The big data supervisor module faces the big data supervisor as a user through the big data supervisor interface.

[0011] It is used to summarize the reports uploaded by various report issuing departments, extract information from the reports, analyze data and standardize them to obtain the ledger, and upload the ledger to the report reporting module. The ledger includes common ledgers and personalized characteristics of each department. The common ledger is a ledger generated based on the common information of each report, and the personalized ledger is a ledger generated by integrating the information in each report that does not belong to the common ledger;

[0012] The report submission module is directed to the report submission departments as users through the report submission interface, and is used to support each report submission department to review and confirm the issued ledgers. For the confirmed ledgers, it is used to support the report submission departments to fill in the ledgers. For the ledgers that have not passed the confirmation, it is used to return the modification opinions fed back by the report submission departments to the big data supervisor module; correspondingly, the big data supervisor module is used to support the big data supervisory department to modify the ledgers based on the modification opinions.

[0013] As a preferred option, the common ledger covers six categories: people, places, events, things, emotions and organizations, and is used to comprehensively record and track all types of information related to government affairs. The common ledger is jointly maintained by all report-submitting departments;

[0014] Personalized ledgers focus on specific government affairs, including the following types:

[0015] Department-specific ledgers: ledgers customized for specific government affairs, including tax registration and tax statistics ledgers, environmental monitoring and pollution treatment record ledgers;

[0016] Project tracking ledger: used to record project progress, fund usage and effect evaluation;

[0017] Emergency management ledger: A ledger set up for emergencies or emergency situations to record all aspects of emergency response, resource allocation and rescue operations;

[0018] Livelihood service records: For social security, health care, and education, used to record service provision and public feedback;

[0019] Management records of specific groups: management and service records for low-income families, the disabled and the elderly.

[0020] Preferably, the ledger is obtained by extracting information, analyzing data and standardizing the report, including the following steps:

[0021] Convert the report format to obtain a processable structured data format. For table reports, extract the data in the table. For unstructured table texts with fixed template styles, convert the table text format to a structured data format based on computer vision technology and optical character recognition technology.

[0022] For reports in structured data formats, the information extracted from the reports is fused and deduplicated. Through fusion and deduplication operations, the same information can be described uniformly in different reports, and redundant information can be deduplicated to achieve consistency and uniqueness of the information.

[0023] For the information after fusion and deduplication, classify it according to its relevance and the category of the report to which it belongs;

[0024] A ledger is constructed based on the classified information, which includes key data points, including time, place, organization, policy name, institution name and personnel name.

[0025] Preferably, for unstructured table text with a fixed template style, the table text format is converted into a structured data format through the following steps:

[0026] Preprocessing: Get historical reports as sample reports, convert PDF documents or Word documents with fixed template styles into images, and perform page segmentation, denoising, and image enhancement operations on the documents;

[0027] Model building: Building a layout analysis model based on computer vision technology, building a table recognition and positioning model based on the Canny edge detection algorithm, and building a text extraction model based on optical character recognition technology. The layout analysis model is used to perform layout analysis on the input image. The table recognition and positioning model is used to identify and locate elements in the table, including rows, columns, and cell boundaries. The text extraction model is used to identify and extract text information in the table and convert the text in the image into an editable text format.

[0028] Model training: Based on the layout analysis model, table recognition and positioning model, and text extraction model, the pre-processed sample report data is structured and processed, and the information extracted from the report is converted into structured data. Based on the structured data, the layout analysis model, table recognition and positioning model, and text extraction model are trained to obtain the trained layout analysis model, table recognition and positioning model, and text extraction model;

[0029] Generate structured data: For the report to be processed, pre-process the report, and perform data structure post-processing on the pre-processed report based on the trained layout analysis model, table recognition and positioning model, and text extraction model, and convert the information extracted from the report to be processed into structured data;

[0030] Data verification and correction: Verify and correct the generated structured data, and optimize the trained layout analysis model, table recognition and positioning model, and text extraction model based on feedback.

[0031] Preferably, when information extracted from reports is integrated and deduplicated, the information is treated as an entity, and a relationship extraction algorithm is used to capture the associations between entities and to construct logical links between information. Based on the associations between entities, logical links between information, and a pre-configured government domain knowledge base, a unified description of the same information in different reports is achieved, and redundant information is deduplicated.

[0032] In a second aspect, the present invention provides a grassroots report fusion method, which implements grassroots report fusion through a grassroots report fusion system as described in any one of the first aspects, and the method includes the following steps:

[0033] Report issuance: Each report issuance department uploads the daily collected data through the report issuance module, and uploads the data to the big data supervisor module in the form of reports;

[0034] Ledger generation: The big data management department uses the big data management module to aggregate reports uploaded by various report-issuing departments, extract information from the reports, perform data analysis, and standardize the reports to obtain ledgers, and then upload the ledgers to the report submission module. The ledgers include common ledgers and personalized features of each department. Common ledgers are ledgers generated based on the common information of each report, and personalized ledgers are ledgers generated by integrating information in each report that does not belong to the common ledger.

[0035] Report submission: Each report submission department reviews and confirms the issued ledgers through the report submission module. For the confirmed ledgers, the report submission department fills in the ledgers. For the ledgers that fail to pass the confirmation, the modification opinions fed back by the report submission department are returned to the big data supervisor module; correspondingly, the big data supervisory department modifies the ledgers through the big data supervisor module based on the modification opinions.

[0036] As a preferred option, the common ledger covers six categories: people, places, events, things, emotions and organizations, and is used to comprehensively record and track all types of information related to government affairs. The common ledger is jointly maintained by all report-submitting departments;

[0037] Personalized ledgers focus on specific government affairs, including the following types:

[0038] Department-specific ledgers: ledgers customized for specific government affairs, including tax registration and tax statistics ledgers, environmental monitoring and pollution treatment record ledgers;

[0039] Project tracking ledger: used to record project progress, fund usage and effect evaluation;

[0040] Emergency management ledger: A ledger set up for emergencies or emergency situations to record all aspects of emergency response, resource allocation and rescue operations;

[0041] Livelihood service records: For social security, health care, and education, used to record service provision and public feedback;

[0042] Management records of specific groups: management and service records for low-income families, the disabled and the elderly.

[0043] Preferably, the ledger is obtained by extracting information, analyzing data and standardizing the report, including the following steps:

[0044] Convert the report format to obtain a processable structured data format. For table reports, extract the data in the table. For unstructured table texts with fixed template styles, convert the table text format to a structured data format based on computer vision technology and optical character recognition technology.

[0045] For reports in structured data formats, the information extracted from the reports is fused and deduplicated. Through fusion and deduplication operations, the same information can be described uniformly in different reports, and redundant information can be deduplicated to achieve consistency and uniqueness of the information.

[0046] For the information after fusion and deduplication, classify it according to its relevance and the category of the report to which it belongs;

[0047] A ledger is constructed based on the classified information, which includes key data points, including time, place, organization, policy name, institution name and personnel name.

[0048] Preferably, for unstructured table text with a fixed template style, the table text format is converted into a structured data format through the following steps:

[0049] Preprocessing: Get historical reports as sample reports, convert PDF documents or Word documents with fixed template styles into images, and perform page segmentation, denoising, and image enhancement operations on the documents;

[0050] Model building: Building a layout analysis model based on computer vision technology, building a table recognition and positioning model based on the Canny edge detection algorithm, and building a text extraction model based on optical character recognition technology. The layout analysis model is used to perform layout analysis on the input image. The table recognition and positioning model is used to identify and locate elements in the table, including rows, columns, and cell boundaries. The text extraction model is used to identify and extract text information in the table and convert the text in the image into an editable text format.

[0051] Model training: Based on the layout analysis model, table recognition and positioning model, and text extraction model, the pre-processed sample report data is structured and processed, and the information extracted from the report is converted into structured data. Based on the structured data, the layout analysis model, table recognition and positioning model, and text extraction model are trained to obtain the trained layout analysis model, table recognition and positioning model, and text extraction model;

[0052] Generate structured data: For the report to be processed, pre-process the report, and perform data structure post-processing on the pre-processed report based on the trained layout analysis model, table recognition and positioning model, and text extraction model, and convert the information extracted from the report to be processed into structured data;

[0053] Data verification and correction: Verify and correct the generated structured data, and optimize the trained layout analysis model, table recognition and positioning model, and text extraction model based on feedback.

[0054] Preferably, when information extracted from reports is integrated and deduplicated, the information is treated as an entity, and a relationship extraction algorithm is used to capture the associations between entities and to construct logical links between information. Based on the associations between entities, logical links between information, and a pre-configured government domain knowledge base, a unified description of the same information in different reports is achieved, and redundant information is deduplicated.

[0055] The grassroots report fusion system and method of the present invention have the following advantages:

[0056] 1. It realizes the automatic filling of data reports, reducing the repetitive work of grassroots staff in data entry, verification, and sorting;

[0057] 2. The data sharing mechanism breaks down the barriers between traditional departments, allowing each department to make decisions and take actions based on the same set of data, thereby achieving effective allocation and efficient use of resources;

[0058] 3. Use computer vision technology, Canny edge detection algorithm and optical character recognition technology to conduct in-depth analysis and verification of raw data, thereby ensuring the accuracy and consistency of the data, providing a reliable basis for grassroots governance, reducing problems and risks caused by human errors, and improving the scientificity and accuracy of decision-making;

[0059] 4. It significantly reduces the work pressure of grassroots staff and improves the efficiency and quality of data processing, providing strong support for the modernization and scientificization of grassroots governance. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only 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.

[0061] The present invention is further described below in conjunction with the accompanying drawings.

[0062] Figure 1 This is a flowchart of a grassroots report fusion method in Example 2. DETAILED DESCRIPTION

[0063] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it. However, the embodiments are not intended to limit the present invention. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments may be combined with each other.

[0064] The embodiments of the present invention provide a basic report fusion system and method for solving the technical problem of how to optimize data collection and fusion processes and provide data reports efficiently and accurately.

[0065] Embodiment 1:

[0066] The present invention discloses a grassroots report fusion system, which comprises a report issuing module, a big data management module and a report submitting module.

[0067] The report issuing module is used to support each report issuing department to upload daily collected data and upload the data to the big data supervisor module in the form of reports through the report issuing interface and the report issuing department as a user.

[0068] The report distribution module of this embodiment supports each department to upload reports in different formats, such as Excel, PDF, Word, CSV, etc.

[0069] The big data supervisor module is oriented to the big data supervisory department as a user through the big data supervisor interface. It is used to summarize the reports uploaded by the various report issuing departments, obtain the ledger by extracting information, analyzing data and standardizing the reports, and upload the ledger to the report reporting module. The ledger includes common ledgers and personalized characteristics of each department. The common ledger is a ledger generated based on the common information of each report, and the personalized ledger is a ledger generated by integrating the information in each report that does not belong to the common ledger.

[0070] In this embodiment, the common ledger covers six categories, namely people, places, things, objects, emotions, and organizations, and is used to comprehensively and systematically record and track various types of information related to government work. This ledger is designed to meet the basic needs of daily government management and decision-making support. The common ledger is jointly maintained by various departments, reducing the duplication of issuance by various departments.

[0071] Compared with common ledgers, personalized ledgers focus more on meeting the special needs of specific government departments, specific policies or projects. Personalized ledgers are more targeted and practical, and can help government departments better perform their duties, improve work efficiency, optimize resource allocation, and thus better serve the public. These ledgers are usually updated and optimized regularly according to actual needs to adapt to the continuous changes and development of government work. These personalized ledgers may include but are not limited to the following types:

[0072] (1) Department-specific ledgers: ledgers customized for the functions and work characteristics of specific government departments. For example, the tax department may have a special tax registration and tax statistics ledger, and the environmental protection department may have an environmental monitoring and pollution treatment record ledger, etc.

[0073] (2) Project tracking records: records designed for the implementation of specific government projects or policies. These records will record in detail information such as project progress, fund use, and effect evaluation, which will help to comprehensively monitor and manage the project.

[0074] (3) Emergency management records: records set up for emergencies or emergencies, recording all aspects of emergency response, resource allocation, rescue operations, and other information to improve the government's ability to respond to emergencies;

[0075] (4) Livelihood service records: For livelihood services such as social security, health care, and education, record service provision, public feedback, and other information to optimize service quality and efficiency;

[0076] (5) Management records for specific groups: such as management and service records for specific groups such as low-income families, the disabled, and the elderly, in order to provide more accurate and personalized services.

[0077] In this embodiment, the report fusion process uses intelligent algorithms and data analysis technology to perform preliminary sorting, key information extraction, data analysis and standardization on various types of collected reports, and finally automatically generates a simplified ledger.

[0078] As a specific implementation, the ledger is obtained by extracting information, analyzing data and standardizing the report, including the following steps:

[0079] (1) Converting the report format to obtain a processable structured data format, wherein for table reports, extracting data from the table; for unstructured table texts in fixed template styles, converting the table text format to a structured data format based on computer vision technology and optical character recognition technology;

[0080] (2) For reports in structured data formats, the information extracted from the reports is fused and deduplicated. Through the fusion and deduplication operation, the same information is described uniformly in different reports, and redundant information is removed to achieve consistency and uniqueness of the information.

[0081] (3) Classify the information after fusion and deduplication based on its relevance and the category of the report to which it belongs;

[0082] (4) Construct a ledger based on the classified information, which includes key data points, including time, location, organization, policy name, agency name, and personnel name.

[0083] Among them, for unstructured table text with a fixed template style, the table text format is converted into a structured data format through the following steps:

[0084] (1) Preprocessing: Obtain historical reports as sample reports. For PDF documents or Word documents with fixed template styles, convert the documents into images and perform page segmentation, denoising, and image enhancement operations on the documents.

[0085] (2) Model construction: construct a layout analysis model based on computer vision technology, construct a table recognition and positioning model based on the Canny edge detection algorithm, and construct a text extraction model based on optical character recognition technology. The layout analysis model is used to perform layout analysis on the input image. The table recognition and positioning model is used to identify and locate elements in the table, including rows, columns, and cell boundaries. The text extraction model is used to identify and extract text information in the table and convert the text in the image into an editable text format.

[0086] (3) Model training: The pre-processed sample report data is structured based on the layout analysis model, table recognition and positioning model, and text extraction model, and the information extracted from the report is converted into structured data. The layout analysis model, table recognition and positioning model, and text extraction model are trained based on the structured data to obtain the trained layout analysis model, table recognition and positioning model, and text extraction model;

[0087] (4) Generating structured data: For the report to be processed, pre-process the report, and perform data structure post-processing on the pre-processed report based on the trained layout analysis model, table recognition and positioning model, and text extraction model, so as to convert the information extracted from the report to be processed into structured data;

[0088] (5) Data verification and correction: Verify and correct the generated structured data, and optimize the trained layout analysis model, table recognition and positioning model, and text extraction model based on the feedback.

[0089] Among them, when the information extracted from the report is integrated and deduplicated, the information is taken as an entity, and the relationship extraction algorithm is used to capture the association between entities, such as "policy-issuing agency", "policy-implementation time", "agency-person in charge", etc., and build logical links between information. Based on the association between entities, logical links between information and the pre-configured government affairs knowledge base, a unified description of the same information in different reports is achieved, and redundant information is removed to improve the accuracy and efficiency of relationship extraction.

[0090] In this embodiment, the ledger contains all key data points, such as time, location, organization, policy name, institution name, personnel name, etc. For repetitive data, the only source will be confirmed. For example, data related to people (name, gender, ID number, address, etc.) will be based on the data provided by the public security department, and data related to houses (house address, building area, service life, building form, etc.) will be based on the data provided by the housing and construction department. At the same time, the source report and source department of each field will be recorded to ensure the traceability of the data.

[0091] The report submission module is directed to the report submission departments as users through the report submission interface, and is used to support each report submission department to review and confirm the issued ledgers. For the ledgers that have passed the confirmation, it is used to support the report submission departments to fill in the ledgers. For the ledgers that have not passed the confirmation, it is used to return the modification opinions fed back by the report submission departments to the big data supervisor module; correspondingly, the big data supervisor module is used to support the big data supervisory department to modify the ledgers based on the modification opinions.

[0092] The system of this embodiment optimizes the data collection and utilization process, realizes one-time data collection and effective application in many aspects. By improving the digitalization and intelligence level of grassroots governance, the system is committed to eliminating unnecessary work. Based on this system, grassroots cadres will no longer need to fill out forms repeatedly or report the same data to multiple departments, thereby freeing up more time to conduct field research at the grassroots level and being able to focus more on the execution and implementation of work.

[0093] Embodiment 2:

[0094] This embodiment provides a grassroots report fusion method, which implements grassroots report fusion through the system disclosed in Example 1, including three steps: report issuance, ledger generation and report submission.

[0095] Step S100: Report issuance: Each report issuance department uploads the daily collected data through the report issuance module, and uploads the data in the form of reports to the big data supervisor module.

[0096] The report distribution in this embodiment supports each department to upload reports in different formats, such as Excel, PDF, Word, CSV, etc.

[0097] Step S200 ledger generation: The big data management department summarizes the reports uploaded by the various report issuing departments through the big data management module, obtains the ledger by extracting information, analyzing data and standardizing the reports, and uploads the ledger to the report reporting module, where the ledger includes common ledgers and personalized characteristics of each department. The common ledger is a ledger generated based on the common information of each report, and the personalized ledger is a ledger generated by integrating the information in each report that does not belong to the common ledger.

[0098] In this embodiment, the common ledger covers six categories, namely people, places, things, objects, emotions, and organizations, and is used to comprehensively and systematically record and track various types of information related to government work. This ledger is designed to meet the basic needs of daily government management and decision-making support. The common ledger is jointly maintained by various departments, reducing the duplication of issuance by various departments.

[0099] Compared with common ledgers, personalized ledgers focus more on meeting the special needs of specific government departments, specific policies or projects. Personalized ledgers are more targeted and practical, and can help government departments better perform their duties, improve work efficiency, optimize resource allocation, and thus better serve the public. These ledgers are usually updated and optimized regularly according to actual needs to adapt to the continuous changes and development of government work. These personalized ledgers may include but are not limited to the following types:

[0100] (1) Department-specific ledgers: ledgers customized for the functions and work characteristics of specific government departments. For example, the tax department may have a special tax registration and tax statistics ledger, and the environmental protection department may have an environmental monitoring and pollution treatment record ledger, etc.

[0101] (2) Project tracking records: records designed for the implementation of specific government projects or policies. These records will record in detail information such as project progress, fund use, and effect evaluation, which will help to comprehensively monitor and manage the project.

[0102] (3) Emergency management records: records set up for emergencies or emergencies, recording all aspects of emergency response, resource allocation, rescue operations, and other information to improve the government's ability to respond to emergencies;

[0103] (4) Livelihood service records: For livelihood services such as social security, health care, and education, record service provision, public feedback, and other information to optimize service quality and efficiency;

[0104] (5) Management records for specific groups: such as management and service records for specific groups such as low-income families, the disabled, and the elderly, in order to provide more accurate and personalized services.

[0105] In this embodiment, the report fusion process uses intelligent algorithms and data analysis technology to perform preliminary sorting, key information extraction, data analysis and standardization on various types of collected reports, and finally automatically generates a simplified ledger.

[0106] As a specific implementation, the ledger is obtained by extracting information, analyzing data and standardizing the report, including the following steps:

[0107] (1) Converting the report format to obtain a processable structured data format, wherein for table reports, extracting data from the table; for unstructured table texts in fixed template styles, converting the table text format to a structured data format based on computer vision technology and optical character recognition technology;

[0108] (2) For reports in structured data formats, the information extracted from the reports is fused and deduplicated. Through the fusion and deduplication operation, the same information is described uniformly in different reports, and redundant information is removed to achieve consistency and uniqueness of the information.

[0109] (3) Classify the information after fusion and deduplication based on its relevance and the category of the report to which it belongs;

[0110] (4) Construct a ledger based on the classified information, which includes key data points, including time, location, organization, policy name, agency name, and personnel name.

[0111] Among them, for unstructured table text with a fixed template style, the table text format is converted into a structured data format through the following steps:

[0112] (1) Preprocessing: Obtain historical reports as sample reports. For PDF documents or Word documents with fixed template styles, convert the documents into images and perform page segmentation, denoising, and image enhancement operations on the documents.

[0113] (2) Model construction: construct a layout analysis model based on computer vision technology, construct a table recognition and positioning model based on the Canny edge detection algorithm, and construct a text extraction model based on optical character recognition technology. The layout analysis model is used to perform layout analysis on the input image. The table recognition and positioning model is used to identify and locate elements in the table, including rows, columns, and cell boundaries. The text extraction model is used to identify and extract text information in the table and convert the text in the image into an editable text format.

[0114] (3) Model training: The pre-processed sample report data is structured based on the layout analysis model, table recognition and positioning model, and text extraction model, and the information extracted from the report is converted into structured data. The layout analysis model, table recognition and positioning model, and text extraction model are trained based on the structured data to obtain the trained layout analysis model, table recognition and positioning model, and text extraction model;

[0115] (4) Generating structured data: For the report to be processed, pre-process the report, and perform data structure post-processing on the pre-processed report based on the trained layout analysis model, table recognition and positioning model, and text extraction model, so as to convert the information extracted from the report to be processed into structured data;

[0116] (5) Data verification and correction: Verify and correct the generated structured data, and optimize the trained layout analysis model, table recognition and positioning model, and text extraction model based on the feedback.

[0117] Among them, when the information extracted from the report is integrated and deduplicated, the information is taken as an entity, and the relationship extraction algorithm is used to capture the association between entities, such as "policy-issuing agency", "policy-implementation time", "agency-person in charge", etc., and build logical links between information. Based on the association between entities, logical links between information and the pre-configured government affairs knowledge base, a unified description of the same information in different reports is achieved, and redundant information is removed to improve the accuracy and efficiency of relationship extraction.

[0118] In this embodiment, the ledger contains all key data points, such as time, location, organization, policy name, institution name, personnel name, etc. For repetitive data, the only source will be confirmed. For example, data related to people (name, gender, ID number, address, etc.) will be based on the data provided by the public security department, and data related to houses (house address, building area, service life, building form, etc.) will be based on the data provided by the housing and construction department. At the same time, the source report and source department of each field will be recorded to ensure the traceability of the data.

[0119] Step S300 report submission: each report submission department reviews and confirms the issued ledger through the report submission module. For the confirmed ledger, the report submission department fills in the ledger. For the ledger that fails to pass the confirmation, the modification opinions fed back by the report submission department are returned to the big data supervisor module; correspondingly, the big data supervisory department modifies the ledger based on the modification opinions through the big data supervisor module.

[0120] The method of this embodiment realizes one-time data collection and effective application in multiple aspects. Grassroots staff will no longer need to fill out forms repeatedly or report the same data to multiple departments, which reduces the workload and improves work efficiency.

[0121] The present invention has been shown and described in detail above through the accompanying drawings and preferred embodiments. However, the present invention is not limited to these disclosed embodiments. Based on the above multiple embodiments, those skilled in the art can know that the means in the above different embodiments can be combined to obtain more embodiments of the present invention, and these embodiments are also within the protection scope of the present invention.

Claims

1. A grassroots report fusion system, characterized in that: It includes report issuing module, big data supervisor module and report submitting module; The report issuing module communicates with the report issuing department as a user through the report issuing interface, and is used to support each report issuing department to upload the daily collected data, and upload the data to the big data supervisor module in the form of reports; The big data supervisor module is used to aggregate reports uploaded by various report issuing departments through the big data supervisor interface, extract information from the reports, perform data analysis and standardization processing to obtain a ledger, and upload the ledger to the report reporting module, wherein the ledger includes a common ledger and personalized characteristics of each department. The common ledger is a ledger generated based on the common information of each report, and the personalized ledger is a ledger generated by integrating the information in each report that does not belong to the common ledger; The report submission module is directed to the report submission departments as users through the report submission interface, and is used to support each report submission department to review and confirm the issued ledgers. For the confirmed ledgers, it is used to support the report submission departments to fill in the ledgers. For the ledgers that have not passed the confirmation, it is used to return the modification opinions fed back by the report submission departments to the big data supervisor module; correspondingly, the big data supervisor module is used to support the big data supervisory department to modify the ledgers based on the modification opinions.

2. The grassroots report fusion system according to claim 1 is characterized in that: The common ledger covers six categories: people, places, events, things, emotions and organizations. It is used to comprehensively record and track all kinds of information related to government affairs. The common ledger is jointly maintained by various report submission departments. Personalized ledgers focus on specific government affairs, including the following types: Department-specific ledgers: ledgers customized for specific government affairs, including tax registration and tax statistics ledgers, environmental monitoring and pollution treatment record ledgers; Project tracking ledger: used to record project progress, fund usage and effect evaluation; Emergency management ledger: A ledger set up for emergencies or emergency situations to record all aspects of emergency response, resource allocation and rescue operations; Livelihood service records: For social security, health care, and education, used to record service provision and public feedback; Management records of specific groups: management and service records for low-income families, the disabled and the elderly.

3. The grassroots report fusion system according to claim 1 or 2, characterized in that: The ledger is obtained by extracting information, analyzing data and standardizing the report, including the following steps: Convert the report format to obtain a processable structured data format. For table reports, extract the data in the table. For unstructured table texts with fixed template styles, convert the table text format to a structured data format based on computer vision technology and optical character recognition technology. For reports in structured data formats, the information extracted from the reports is fused and deduplicated. Through fusion and deduplication operations, the same information can be described uniformly in different reports, and redundant information can be deduplicated to achieve consistency and uniqueness of the information. For the information after fusion and deduplication, classify it according to its relevance and the category of the report to which it belongs; A ledger is constructed based on the classified information, which includes key data points, including time, place, organization, policy name, institution name and personnel name.

4. The grassroots report fusion system according to claim 3 is characterized in that: For unstructured table text with a fixed template style, convert the table text format into a structured data format through the following steps: Preprocessing: Get historical reports as sample reports, convert PDF documents or Word documents with fixed template styles into images, and perform page segmentation, denoising, and image enhancement operations on the documents; Model building: Building a layout analysis model based on computer vision technology, building a table recognition and positioning model based on the Canny edge detection algorithm, and building a text extraction model based on optical character recognition technology. The layout analysis model is used to perform layout analysis on the input image. The table recognition and positioning model is used to identify and locate elements in the table, including rows, columns, and cell boundaries. The text extraction model is used to identify and extract text information in the table and convert the text in the image into an editable text format. Model training: Based on the layout analysis model, table recognition and positioning model, and text extraction model, the pre-processed sample report data is structured and processed, and the information extracted from the report is converted into structured data. Based on the structured data, the layout analysis model, table recognition and positioning model, and text extraction model are trained to obtain the trained layout analysis model, table recognition and positioning model, and text extraction model; Generate structured data: For the report to be processed, pre-process the report, and perform data structure post-processing on the pre-processed report based on the trained layout analysis model, table recognition and positioning model, and text extraction model, and convert the information extracted from the report to be processed into structured data; Data verification and correction: Verify and correct the generated structured data, and optimize the trained layout analysis model, table recognition and positioning model, and text extraction model based on feedback.

5. The grassroots report fusion system according to claim 3 is characterized in that: When integrating and deduplicating information extracted from reports, the information is treated as an entity, and the relationship extraction algorithm is used to capture the associations between entities and build logical links between information. Based on the associations between entities, logical links between information, and the pre-configured government domain knowledge base, a unified description of the same information in different reports is achieved, and redundant information is deduplicated.

6. A grassroots report fusion method, characterized in that: The grassroots report fusion is realized by a grassroots report fusion system as described in any one of claims 1 to 5, and the method comprises the following steps: Report issuance: Each report issuance department uploads the daily collected data through the report issuance module, and uploads the data to the big data supervisor module in the form of reports; Ledger generation: The big data management department uses the big data management module to aggregate reports uploaded by various report-issuing departments, extract information from the reports, perform data analysis, and standardize the reports to obtain ledgers, and then upload the ledgers to the report submission module. The ledgers include common ledgers and personalized features of each department. Common ledgers are ledgers generated based on the common information of each report, and personalized ledgers are ledgers generated by integrating information in each report that does not belong to the common ledger. Report submission: Each report submission department reviews and confirms the issued ledgers through the report submission module. For the confirmed ledgers, the report submission department fills in the ledgers. For the ledgers that fail to pass the confirmation, the modification opinions fed back by the report submission department are returned to the big data supervisor module; correspondingly, the big data supervisory department modifies the ledgers through the big data supervisor module based on the modification opinions.

7. The grassroots report fusion method according to claim 6 is characterized in that: The common ledger covers six categories: people, places, events, things, emotions and organizations. It is used to comprehensively record and track all kinds of information related to government affairs. The common ledger is jointly maintained by various report submission departments. Personalized ledgers focus on specific government affairs, including the following types: Department-specific ledgers: ledgers customized for specific government affairs, including tax registration and tax statistics ledgers, environmental monitoring and pollution treatment record ledgers; Project tracking ledger: used to record project progress, fund usage and effect evaluation; Emergency management ledger: A ledger set up for emergencies or emergency situations to record all aspects of emergency response, resource allocation and rescue operations; Livelihood service records: For social security, health care, and education, used to record service provision and public feedback; Management records of specific groups: management and service records for low-income families, the disabled and the elderly.

8. The grassroots report fusion method according to claim 6 or 7, characterized in that: The ledger is obtained by extracting information, analyzing data and standardizing the report, including the following steps: Convert the report format to obtain a processable structured data format. For table reports, extract the data in the table. For unstructured table texts with fixed template styles, convert the table text format to a structured data format based on computer vision technology and optical character recognition technology. For reports in structured data formats, the information extracted from the reports is fused and deduplicated. Through fusion and deduplication operations, the same information can be described uniformly in different reports, and redundant information can be deduplicated to achieve consistency and uniqueness of the information. For the information after fusion and deduplication, classify it according to its relevance and the category of the report to which it belongs; A ledger is constructed based on the classified information, which includes key data points, including time, place, organization, policy name, institution name and personnel name.

9. The grassroots report fusion method according to claim 8 is characterized in that: For unstructured table text with a fixed template style, convert the table text format into a structured data format through the following steps: Preprocessing: Get historical reports as sample reports, convert PDF documents or Word documents with fixed template styles into images, and perform page segmentation, denoising, and image enhancement operations on the documents; Model building: Building a layout analysis model based on computer vision technology, building a table recognition and positioning model based on the Canny edge detection algorithm, and building a text extraction model based on optical character recognition technology. The layout analysis model is used to perform layout analysis on the input image. The table recognition and positioning model is used to identify and locate elements in the table, including rows, columns, and cell boundaries. The text extraction model is used to identify and extract text information in the table and convert the text in the image into an editable text format. Model training: Based on the layout analysis model, table recognition and positioning model, and text extraction model, the pre-processed sample report data is structured and processed, and the information extracted from the report is converted into structured data. Based on the structured data, the layout analysis model, table recognition and positioning model, and text extraction model are trained to obtain the trained layout analysis model, table recognition and positioning model, and text extraction model; Generate structured data: For the report to be processed, pre-process the report, and perform data structure post-processing on the pre-processed report based on the trained layout analysis model, table recognition and positioning model, and text extraction model, and convert the information extracted from the report to be processed into structured data; Data verification and correction: Verify and correct the generated structured data, and optimize the trained layout analysis model, table recognition and positioning model, and text extraction model based on feedback.

10. The grassroots report fusion method according to claim 8, characterized in that: When integrating and deduplicating information extracted from reports, the information is treated as an entity, and the relationship extraction algorithm is used to capture the associations between entities and build logical links between information. Based on the associations between entities, logical links between information, and the pre-configured government domain knowledge base, a unified description of the same information in different reports is achieved, and redundant information is deduplicated.

Images