Government affair large model-based machine account intelligent similarity matching generation method and system in urban grassroots government

Abstract

The invention discloses an intelligent machine account similarity matching generation method and system in urban grassroots government based on a government affair large model, belongs to the technical field of large model and urban management, and aims to solve the technical problem of how to automatically input report data, realize automatic matching of machine account metadata and report metadata, promote data standardization and improve the efficiency of data processing. According to the technical scheme, the method comprises the following steps: data integration and preprocessing: collecting various kinds of original data related to urban government through a government affair cloud platform, and carrying out cleaning and duplicate removal preprocessing operation on the original data to obtain preprocessed original data; meanwhile, personal or sensitive information is processed by adopting an encryption technology and a desensitization technology, so that the data security and compliance are ensured; feature extraction and large model construction; and carrying out intelligent similarity matching on the ledgers.

Description

Technical Field

[0001] The present invention relates to the field of big models and urban management technology, and specifically to a method and system for generating intelligent similarity matching of ledgers in urban grassroots governance based on a big government model. Background Art

[0002] With the acceleration of urbanization, urban grassroots governance faces increasingly complex challenges. Although there are some automation tools in existing technologies, they often lack sufficient intelligent support and are difficult to cope with the complex and changing needs of urban management.

[0003] Traditional ledger management relies heavily on manual operations, which is not only inefficient, but also often leads to errors such as multiple reporting, repeated requests for data, and poor information interactivity and sharing, resulting in information asymmetry and affecting urban management effectiveness.

[0004] The reporting system is scattered in various business lines, with multiple entries of the same content, repeated reporting, and inconsistent formats. Various types of data are scattered in business systems at different levels in the city, making it difficult to share data. There are many reports, many lines involved, many repeated reports, and a lot of time. Reports are not submitted in time, report data is inaccurate, data coverage is not comprehensive, and the task of merging tables is heavy. Data collection tasks are heavy and difficult to update, there are many data silos and it is difficult to share, data is scattered and difficult to integrate, and data standards are not standardized and difficult to call.

[0005] Therefore, how to automatically input report data, automatically match ledger metadata and report metadata, promote data standardization, and improve work efficiency and data accuracy is a technical problem that needs to be solved urgently. Summary of the invention

[0006] The technical task of the present invention is to provide a method and system for generating intelligent similarity matching of ledgers in urban grassroots governance based on a big government model, so as to solve the problem of how to automatically enter report data, realize automatic matching of ledger metadata and report metadata, promote data standardization, and improve work efficiency and data accuracy.

[0007] The technical task of the present invention is achieved in the following way: a method for generating intelligent similarity matching of ledgers in urban grassroots governance based on a government affairs big model, the method is as follows:

[0008] Data integration and preprocessing: Collect all kinds of raw data related to urban governance through the government cloud platform, and perform preprocessing operations such as cleaning and deduplication on the raw data to obtain the preprocessed raw data; at the same time, use encryption technology or desensitization technology to process personal or sensitive information to ensure data security and compliance; among them, the raw data related to various types of urban governance include natural person information, real estate registration information, industrial and commercial registration information, urban planning information, and public safety record information;

[0009] Feature extraction and large model construction: Based on government big data, key features that are highly correlated with the matching degree of the ledger are extracted, and the large language model Transformer is used to build a matching government big model;

[0010] Intelligent similarity matching of ledgers: Through context-aware matching of ledgers, complex contextual information is converted into a format that can be understood by large modules, and accurate time screening and priority sorting are performed according to specific conditions to improve governance efficiency and accuracy; then, through case learning and pattern recognition of patterns and rules, the matching large model is iteratively improved, the knowledge obtained from pattern recognition is re-matched for similarity, and updated and learned in real time.

[0011] As a preferred method, feature extraction and large model construction are as follows:

[0012] Feature extraction: In the context of government big data, identify and extract the most helpful information for a specific task or analysis in the data set, i.e., key features; key features include location information, time series, time type, and subject attributes;

[0013] Model selection: The selection of government big data models is based on data characteristics and matching accuracy requirements. The characteristics of government data are deeply analyzed, and the required matching accuracy is determined according to business needs and goals. High-precision requirements may require more complex government big data models and more training data. In some cases, in order to maintain a balance, a trade-off may be made between accuracy and model complexity. The characteristics of government data include the data structure type (such as text, numbers, time series, etc.), the data noise level, and the data distribution characteristics.

[0014] Model training and optimization: Use historical data sets to train the government affairs model, adjust hyperparameters through cross-validation to ensure the generalization ability of the government affairs model; set the initial parameters of the learning rate, batch size, and number of layers of the government affairs model, and select a suitable loss function and optimizer; then based on the performance results of cross-validation, adjust the hyperparameters of the learning rate, batch size, number of layers, and number of hidden units to achieve the best performance, and use an independent test set for final evaluation to ensure that the government affairs model has good generalization ability and can perform well on new data. Closely observe the loss curve and accuracy indicators during the training process of the government affairs model, look for signs of overfitting or underfitting, and take corresponding measures; then based on the feedback from the test set and business needs, continue to adjust the government affairs model, including increasing or decreasing the number of layers, changing the activation function, and adjusting the regularization strategy.

[0015] Preferably, geographic location: ledger data is usually associated with a specific location, such as the place where the event occurred, the location of the enterprise, and the place where an individual lives;

[0016] Geographical location features are defined by extracting address information, geographic coordinates or area codes, which are helpful for regional analysis; regional analysis includes urban hotspot distribution and regional resource allocation;

[0017] Time series: Time is an important dimension that affects ledger data. The timestamp information of the extracted occurrence time and recording time is used to analyze the time series characteristics of data change trends, periodic fluctuations and timeliness;

[0018] Event type: The event type reflects the nature of the ledger entry. By classifying and marking different event types, data can be grouped and compared to find patterns and associations.

[0019] Subject attributes: Attributes of individuals or organizations related to the ledger. Subject attributes include demographic characteristics, enterprise size, and economic type.

[0020] As a preferred option, the ledger context-aware matching is as follows:

[0021] Context information integration, specifically: first, encode the geographic location information into a format understandable to the government big model; the geographic location information includes geographic coordinates and region codes; second, convert the time information of the event into time series data that can be used in the government big model. The time series data includes timestamps and date differences. The time series data is incorporated into the input of the government big model as an additional feature, or introduced at different levels of the government big model to better capture the spatiotemporal correlation behind the event;

[0022] Conditional screening: Screening events based on geographic location, giving priority to events related to specific locations; specifically: when dealing with urban noise issues, give priority to noise complaints near residential areas or schools; at the same time, consider the time factors of working hours and holidays when the events occur, and screen in the time dimension; during holidays, strengthen supervision of the environment around tourist attractions; in addition, classify and screen events based on event type and urgency to ensure that urgent or high-priority events are given priority;

[0023] Prioritization: Develop context-based rules for event prioritization. Specifically, emergency events take precedence over general events. Learning methods are used to learn event priority patterns based on historical data and automatically adjust priorities. For example, if a certain type of event was generally considered a high priority in the past, it can be automatically set to a high priority in the future. Finally, a comprehensive consideration of the event severity, scope of impact, various attributes of related entities, and contextual information is used to conduct multi-dimensional evaluation and sorting.

[0024] As a preferred method, case study and pattern recognition are as follows:

[0025] Build a comprehensive and detailed case library: systematically collect and organize verified and successfully implemented data matching examples; each case not only includes the final state of successful matching, but also needs to record in detail the key information of input data characteristics, matching strategies adopted and environment configuration during the matching process. Through highly structured data organization, the cases are safely and efficiently stored in advanced database management systems or specially designed knowledge bases, providing a solid foundation for subsequent analysis and learning;

[0026] Use advanced pattern recognition technology to deeply explore the resources in the case library: involving data governance models, aiming to extract hidden rules and patterns from a large number of successfully matched instances, and through reinforcement learning mechanisms, dynamically adjust the model structure based on the results of pattern recognition, such as increasing or reducing the number of layers of the neural network, or changing the branch logic of the decision tree;

[0027] According to the identified effective feature combinations, the feature weights are automatically adjusted to ensure that each match can focus more accurately on the most critical data attributes; based on the identified patterns and rules, the matching government affairs big model is iteratively improved to optimize the structure and parameters of the government affairs big model, and the knowledge obtained from pattern recognition is re-matched, for example, by updating decision rules or optimizing feature weights.

[0028] As a preferred option, the intelligent similarity matching of ledgers also includes real-time updating and learning, as follows:

[0029] Adaptive weight learning: Design an online learning module to dynamically learn the impact of different features on the matching results during the matching process and automatically adjust the feature weights; and use the newly entered data to continuously optimize the matching algorithm and improve the matching accuracy;

[0030] Design online learning module: select appropriate online learning algorithm that can process continuously arriving data and gradually update the government model;

[0031] Determine the appropriate learning rate, control the speed of weight updates, balance the stability of the government model and the need to quickly adapt to new data, and perform dynamic feature weight adjustments, track the impact of different features on matching results, and estimate them through statistical methods or model prediction error analysis;

[0032] According to the results of feature impact analysis, a weight update mechanism is designed to enhance features that contribute greatly to the matching results and suppress features that contribute little;

[0033] Introduce regularization terms in weight updates to avoid overfitting caused by excessive weights, maintain the generalization ability of the government affairs model, and continuously optimize matching to include newly entered data in the training set in real time to ensure that the government affairs model can learn from the latest data in real time;

[0034] Continuously monitor matching accuracy and other relevant performance indicators, promptly detect performance degradation and take action;

[0035] Regularly retrain the government big model based on accumulated new data to absorb more complex patterns and relationships;

[0036] Consider using a big government model fusion strategy to combine the judgments of multiple different big government models and improve the robustness of matching through a weighted or voting mechanism.

[0037] Preferably, the method further includes an early warning feedback mechanism, which is as follows:

[0038] Intelligent early warning and decision support: When the matching results are abnormal or potential problems are found, the early warning is automatically triggered and decision suggestions are provided to the management personnel. The early warning information is promptly notified to the relevant management personnel through the channels of email, SMS or internal message system. At the same time, the learning government affairs big model is used to conduct in-depth analysis of the abnormal causes and provide possible explanations. Performance reports and early warning history records are regularly generated for management personnel to review, and all early warning events and subsequent decision-making operations are recorded to ensure traceability and clear responsibilities.

[0039] Effect evaluation and feedback loop: Establish an evaluation system for matching results, regularly evaluate the performance of the government affairs big model through the indicators of accuracy and recall rate, collect user feedback, continuously iterate and optimize, and build user-friendly feedback channels such as online surveys, user interviews, and customer service records. Quantify user feedback information (such as satisfaction scores) and record it, and analyze it together with the performance indicators of the government affairs big model; conduct in-depth research on specific cases of user feedback to better understand user needs and the performance of the government affairs big model; at the same time, determine the priority and direction of optimization based on performance evaluation and user feedback; after the government affairs big model is updated, closely monitor the performance changes and user feedback of the government affairs big model to ensure that the improvement measures achieve the expected results, regularly review the entire evaluation and feedback process, find potential bottlenecks, and further optimize the process.

[0040] A system for intelligent similarity matching and generating a ledger in urban grassroots governance based on a government affairs big model, the system is used to implement the above-mentioned method for intelligent similarity matching and generating a ledger in urban grassroots governance based on a government affairs big model; the system comprises:

[0041] The data integration and preprocessing module is used to collect various raw data related to urban governance through the government cloud platform, and perform preprocessing operations such as cleaning and deduplication on the raw data to obtain the preprocessed raw data; at the same time, encryption technology or desensitization technology is used to process personal or sensitive information to ensure data security and compliance; among them, various raw data related to urban governance include natural person information, real estate registration information, industrial and commercial registration information, urban planning information, and public safety record information;

[0042] The feature extraction and large model building module is used to extract key features that are highly correlated with the matching degree of the ledger based on the government affairs big data, and use the large language model Transformer to build a matching government affairs big model;

[0043] The ledger intelligent similarity matching module is used to convert complex contextual information into a format that can be understood by large modules through ledger context-aware matching, and perform precise time screening and priority sorting according to specific conditions to improve governance efficiency and accuracy; then, through case learning and pattern recognition of patterns and rules, the matching large model is iteratively improved, the knowledge obtained from pattern recognition is re-matched for similarity, and updated and learned in real time.

[0044] An electronic device comprising: a memory and at least one processor;

[0045] Wherein, the memory stores computer-executable instructions;

[0046] The at least one processor executes the computer execution instructions stored in the memory, so that the at least one processor executes the above-mentioned method for generating intelligent similarity matching of ledgers in urban grassroots governance based on the government affairs big model.

[0047] A computer-readable storage medium having computer execution instructions stored therein. When a processor executes the computer execution instructions, the method for generating intelligent similarity matching of ledgers in urban grassroots governance based on a big government affairs model as described above is implemented.

[0048] The method and system for intelligent similarity matching generation of ledgers in urban grassroots governance based on the government affairs big model of the present invention have the following advantages:

[0049] (1) The present invention improves data processing efficiency: through intelligent similarity matching, similar ledger data entries can be quickly identified and associated, greatly reducing the time for manual review and matching, and improving data processing efficiency;

[0050] (ii) The present invention optimizes resource allocation: similarity matching helps to identify duplicate or similar data records, avoids waste of resources, and ensures that resources are reasonably allocated and effectively utilized;

[0051] (III) The present invention improves data quality: the intelligent matching algorithm can identify and correct errors in the data, ensuring the accuracy of the ledger data, thereby improving the overall data quality;

[0052] (iv) The present invention enhances data monitoring: similarity matching can be used to monitor abnormal or irregular behaviors, for example, in financial management, it can be used to identify potential fraud or errors;

[0053] (V) The present invention supports decision making: providing accurate data support to management and helping to make more scientific and reasonable decisions by analyzing the relationship between similar data;

[0054] (VI) The present invention promotes information sharing: intelligent matching helps to achieve data sharing between different departments or systems, breaking down information silos and enhancing collaborative work capabilities;

[0055] (VII) The present invention reduces operating costs: it reduces the cost waste caused by data errors or duplication of work, and in the long run, it can significantly reduce operating costs;

[0056] (VIII) The present invention enhances risk management: through similarity matching, risks can be better identified and managed, for example, in compliance checks, potential compliance risks can be quickly discovered;

[0057] (IX) This invention promotes digital transformation: Intelligent similarity matching is one of the key technologies for realizing digital management, which helps promote the transformation of organizations to a more efficient and intelligent operation model;

[0058] (10) This invention effectively solves the difficulties faced by grassroots work through the application of technological innovation in grassroots governance, improves grassroots work efficiency and data quality, provides strong support for the modernization of grassroots governance, and has important social value and economic benefits;

[0059] (XI) In order to comprehensively improve the practicality and convenience of users, the present invention uses an intelligent grassroots reporting module to pass the report name and the required account name to the trained and optimized government affairs big model. The government affairs big model analyzes and returns the account names with higher similarity and the corresponding similarity values ​​through the intelligent similarity measurement module. The intelligent grassroots reporting module sorts the accounts from high to low according to the similarity values ​​and displays the accounts to the user. The intelligent grassroots reporting module automatically establishes a corresponding relationship between the account fields and the report fields according to the analysis results of the big model and displays them. After the user confirms that they are correct, he clicks the fill button to automatically complete the filling and generation of the corresponding field data of the report;

[0060] (XII) The present invention realizes intelligent grassroots reports based on the government affairs big model capability, automatically recommends ledgers by report name, and automatically matches ledger metadata with report metadata based on the big model capability after selecting a specific ledger. After confirming the filling, the ledger data is filled into the report; automatically recommends ledgers by report name to avoid multiple entries of the same content, improve work efficiency, break data chimneys, realize data sharing between business systems at different levels, and improve data utilization; automatically fills in data, reduces human errors, and improves data accuracy. Reduce data collection tasks, save time, and improve work efficiency; automatically match ledger metadata with report metadata, promote data standardization, and facilitate data calls. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0062] Attached Figure 1 This is a flowchart of the intelligent similarity matching generation method for ledgers in urban grassroots governance based on the government affairs big model. DETAILED DESCRIPTION

[0063] With reference to the drawings and specific embodiments of the specification, the method and system for generating intelligent similarity matching of ledgers in urban grassroots governance based on the government affairs big model of the present invention are described in detail as follows.

[0064] Embodiment 1:

[0065] As attached Figure 1 As shown, this embodiment provides a method for generating intelligent similarity matching of ledgers in urban grassroots governance based on a government affairs big model, and the method is specifically as follows:

[0066] S1. Data integration and preprocessing: Collect all kinds of raw data related to urban governance through the government cloud platform, and perform preprocessing operations such as cleaning and deduplication on the raw data to obtain the preprocessed raw data; at the same time, encryption technology or desensitization technology is used to process personal or sensitive information to ensure data security and compliance; among them, the raw data related to various types of urban governance include natural person information, real estate registration information, industrial and commercial registration information, urban planning information, and public safety record information;

[0067] S2. Feature extraction and large model construction: Based on government big data, key features that are highly correlated with the matching degree of the ledger are extracted, and a large language model Transformer is used to build a matching government big model;

[0068] S3. Intelligent similarity matching of ledgers: Through ledger context-aware matching, complex context information is converted into a format that can be understood by large modules, and accurate time screening and priority sorting are performed according to specific conditions to improve governance efficiency and accuracy. Then, through case learning and pattern recognition of patterns and rules, the matching large model is iteratively improved, and the knowledge obtained from pattern recognition is re-matched for similarity, and updated and learned in real time.

[0069] S4. Early warning feedback mechanism, as follows:

[0070] S401. Intelligent early warning and decision support: When the matching results are abnormal or potential problems are found, the early warning is automatically triggered and decision suggestions are provided to the management personnel. The early warning information is promptly notified to the relevant management personnel through the channels of email, SMS or internal message system. At the same time, the learning government affairs big model is used to conduct in-depth analysis of the abnormal causes and provide possible explanations. Performance reports and early warning history records are regularly generated for management personnel to review, and all early warning events and subsequent decision-making operations are recorded to ensure traceability and clear responsibilities.

[0071] S402. Effect evaluation and feedback loop: Establish an evaluation system for matching results, regularly evaluate the performance of the government affairs big model through the indicators of accuracy and recall, collect user feedback, continuously iterate and optimize, and build user-friendly feedback channels such as online surveys, user interviews, and customer service records. Quantify user feedback information (such as satisfaction scores) and record it, and analyze it together with the performance indicators of the government affairs big model; conduct in-depth research on specific cases of user feedback to better understand user needs and the performance of the government affairs big model; at the same time, determine the priority and direction of optimization based on performance evaluation and user feedback; after the government affairs big model is updated, closely monitor the performance changes and user feedback of the government affairs big model to ensure that the improvement measures achieve the expected results, regularly review the entire evaluation and feedback process, find potential bottlenecks, and further optimize the process.

[0072] The feature extraction and large model construction in step S2 of this embodiment are specifically as follows:

[0073] S201. Feature extraction: In the context of government big data, identify and extract the most helpful information for a specific task or analysis in the data set, i.e., key features; key features include location information, time series, time type, and subject attributes;

[0074] S202, Model selection: The selection of the government big model is based on data characteristics and matching accuracy requirements. The characteristics of government data are deeply analyzed, and the required matching accuracy is determined according to business needs and goals. High-precision requirements may require more complex government big models and more training data. In some cases, in order to maintain a balance, a trade-off may be made between accuracy and model complexity. The characteristics of government data include the structure type of data (such as text, numbers, time series, etc.), the noise level of the data, and the distribution characteristics of the data.

[0075] S203. Model training and optimization: Use historical data sets to train the government affairs model, adjust hyperparameters through cross-validation to ensure the generalization ability of the government affairs model; set the initial parameters of the learning rate, batch size, and number of layers of the government affairs model, and select a suitable loss function and optimizer; then based on the performance results of cross-validation, adjust the hyperparameters of the learning rate, batch size, number of layers, and number of hidden units to achieve the best performance, and use an independent test set for final evaluation to ensure that the government affairs model has good generalization ability and can perform well on new data, closely observe the loss curve and accuracy indicators during the training process of the government affairs model, look for signs of overfitting or underfitting, and take corresponding measures; then based on the feedback from the test set and business needs, continue to adjust the government affairs model, including increasing or decreasing the number of layers, changing the activation function, and adjusting the regularization strategy.

[0076] The geographical location in step S201 of this embodiment: the ledger data is usually associated with a specific location, which includes the place where the event occurred, the location of the enterprise, and the place of residence of an individual.

[0077] In step S201 of this embodiment, the geographical location feature is defined by extracting address information, geographical coordinates or area codes, which are helpful for regional analysis; the regional analysis includes urban hotspot distribution and regional resource allocation.

[0078] Time series in step S201 of this embodiment: Time is an important dimension that affects the ledger data. The timestamp information of the extracted occurrence time and recording time is used to analyze the time series characteristics of the data's changing trend, periodic fluctuations, and timeliness.

[0079] Event type in step S201 of this embodiment: The event type reflects the nature of the ledger entry. By classifying and marking different event types, data is grouped and compared and analyzed to find patterns and associations.

[0080] In step S201 of this embodiment, subject attributes are attributes of individuals or organizations related to the ledger, and subject attributes include demographic characteristics, enterprise size, and economic type.

[0081] The specific details of the ledger context-aware matching in step S3 of this embodiment are as follows:

[0082] S301, context information integration, specifically: first, encoding the geographic location information into a format understandable to the government affairs big model; the geographic location information includes geographic coordinates and region codes; second, converting the time information of the event into time series data that can be used in the government affairs big model, the time series data includes timestamps and date differences, and the time series data is incorporated into the input of the government affairs big model as an additional feature, or introduced in different levels of the government affairs big model, so as to better capture the spatiotemporal correlation behind the event;

[0083] S302, conditional screening: Screening events based on geographic location, giving priority to events related to specific locations; specifically: when dealing with urban noise issues, giving priority to noise complaints near residential areas or schools; at the same time, considering the time factors of working hours and holidays when the events occur, screening is performed in the time dimension; during holidays, strengthening the supervision of the environment around tourist attractions; in addition, based on the factors of event type and urgency, classifying and screening events to ensure that urgent or high-priority events are given priority;

[0084] S303. Priority sorting: Develop context-based rules for event priority sorting. Specifically, emergency events take precedence over general events. Learning methods are used to learn event priority sorting patterns based on historical data and automatically adjust priorities. For example, if a certain type of event has been generally regarded as a high priority in past processing, it can be automatically set to a high priority in the future. Finally, a comprehensive consideration is given to the severity of the event, the scope of impact, various attributes of the relevant subjects, and contextual information, and multi-dimensional evaluation and sorting is performed.

[0085] The case study and pattern recognition in step S3 of this embodiment are specifically as follows:

[0086] ① Build a comprehensive and detailed case library: systematically collect and organize verified and successfully implemented data matching examples; each case not only includes the final state of successful matching, but also needs to record in detail the key information of input data characteristics, matching strategies adopted and environment configuration during the matching process. Through highly structured data organization, the cases are safely and efficiently stored in advanced database management systems or specially designed knowledge bases, providing a solid foundation for subsequent analysis and learning;

[0087] ② Use advanced pattern recognition technology to deeply explore the resources in the case library: This involves data governance models, which aim to extract hidden rules and patterns from a large number of successfully matched instances, and through reinforcement learning mechanisms, dynamically adjust the model structure based on the results of pattern recognition, such as increasing or reducing the number of layers of the neural network, or changing the branch logic of the decision tree;

[0088] ③ According to the identified effective feature combinations, the feature weights are automatically adjusted to ensure that each match can focus more accurately on the most critical data attributes; based on the identified patterns and rules, the matching government affairs big model is iteratively improved to optimize the structure and parameters of the government affairs big model, and the knowledge obtained from pattern recognition is re-matched, for example, by updating decision rules or optimizing feature weights.

[0089] In step S3 of this embodiment, the intelligent similarity matching of the ledger also includes real-time updating and learning, which are as follows:

[0090] ① Adaptive weight learning: Design an online learning module to dynamically learn the impact of different features on the matching results during the matching process and automatically adjust the feature weights; and use the newly entered data to continuously optimize the matching algorithm and improve the matching accuracy;

[0091] ② Design online learning module: select appropriate online learning algorithm, which can process the continuously arriving data and gradually update the government affairs model;

[0092] ③ Determine the appropriate learning rate, control the speed of weight update, balance the stability of the government model and the need to quickly adapt to new data, and perform dynamic feature weight adjustment, track the impact of different features on matching results, and estimate through statistical methods or model prediction error analysis;

[0093] ④ According to the results of feature impact analysis, design a weight update mechanism to enhance the features that contribute greatly to the matching results and suppress the features that contribute less;

[0094] ⑤ Introduce regularization terms in weight updates to avoid overfitting caused by excessive weights, maintain the generalization ability of the government affairs model, and continuously optimize matching, incorporate newly entered data into the training set in real time, and ensure that the government affairs model can learn from the latest data in real time;

[0095] ⑥Continuously monitor matching accuracy and other relevant performance indicators, promptly detect performance degradation and take measures;

[0096] ⑦ Regularly retrain the government big model based on accumulated new data to absorb more complex patterns and relationships;

[0097] ⑧ Consider using the government big model fusion strategy, combining the judgments of multiple different government big models, and improving the robustness of matching through weighted or voting mechanisms.

[0098] Embodiment 2:

[0099] This embodiment provides a system for intelligent similarity matching generation of urban grassroots governance ledgers based on a government affairs big model, which is used to implement the method for intelligent similarity matching generation of urban grassroots governance ledgers based on a government affairs big model as in Example 1; the system includes:

[0100] The data integration and preprocessing module is used to collect various raw data related to urban governance through the government cloud platform, and perform preprocessing operations such as cleaning and deduplication on the raw data to obtain the preprocessed raw data; at the same time, encryption technology or desensitization technology is used to process personal or sensitive information to ensure data security and compliance; among them, various raw data related to urban governance include natural person information, real estate registration information, industrial and commercial registration information, urban planning information, and public safety record information;

[0101] The feature extraction and large model building module is used to extract key features that are highly correlated with the matching degree of the ledger based on the government affairs big data, and use the large language model Transformer to build a matching government affairs big model;

[0102] The ledger intelligent similarity matching module is used to convert complex contextual information into a format that can be understood by large modules through ledger context-aware matching, and perform precise time screening and priority sorting according to specific conditions to improve governance efficiency and accuracy; then, through case learning and pattern recognition of patterns and rules, the matching large model is iteratively improved, the knowledge obtained from pattern recognition is re-matched for similarity, and updated and learned in real time.

[0103] Embodiment 3:

[0104] This embodiment also provides an electronic device, including: a memory and at least one processor;

[0105] Wherein, the memory stores computer-executable instructions;

[0106] The at least one processor executes the computer execution instructions stored in the memory, so that the at least one processor executes the method for generating intelligent similarity matching of ledgers in urban grassroots governance based on the government affairs big model as described in any one of the present inventions.

[0107] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor may be a microprocessor or any conventional processor, etc.

[0108] The memory can be used to store computer programs and / or modules. The processor realizes various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the terminal, etc. In addition, the memory can also include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SMC), a secure digital (SD) card, a flash memory card, at least one disk storage period, a flash memory device, or other volatile solid-state storage devices.

[0109] Embodiment 4:

[0110] This embodiment also provides a computer-readable storage medium, which stores a plurality of instructions, which are loaded by a processor to enable the processor to execute the method for generating intelligent similarity matching of ledgers in urban grassroots governance based on a government affairs model in any embodiment of the present invention. Specifically, a system or device equipped with a storage medium can be provided, on which software program codes for implementing the functions of any of the above embodiments are stored, and a computer (or CPU or MPU) of the system or device can read and execute the program code stored in the storage medium.

[0111] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.

[0112] The storage medium embodiments for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code can be downloaded from a server computer by a communication network.

[0113] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling an operating system operating on the computer to complete part or all of the actual operations based on instructions from the program code.

[0114] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or written to a memory provided in an expansion unit connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or the expansion unit is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.

[0115] 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 replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent similarity matching generation of ledgers in urban grassroots governance based on a government affairs big model, characterized in that: The method is as follows: Data integration and preprocessing: Collect all kinds of raw data related to urban governance through the government cloud platform, and perform preprocessing operations such as cleaning and deduplication on the raw data to obtain the preprocessed raw data; at the same time, use encryption technology or desensitization technology to process personal or sensitive information to ensure data security and compliance; among them, the raw data related to various types of urban governance include natural person information, real estate registration information, industrial and commercial registration information, urban planning information, and public safety record information; Feature extraction and large model construction: Based on government big data, key features that are highly correlated with the matching degree of the ledger are extracted, and the large language model Transformer is used to build a matching government big model; Intelligent similarity matching of ledgers: Through context-aware matching of ledgers, complex contextual information is converted into a format that can be understood by large modules, and accurate time screening and priority sorting are performed according to specific conditions to improve governance efficiency and accuracy; then, through case learning and pattern recognition of patterns and rules, the matching large model is iteratively improved, the knowledge obtained from pattern recognition is re-matched for similarity, and updated and learned in real time.

2. According to claim 1, the method for generating intelligent similarity matching of ledgers in urban grassroots governance based on the government affairs big model is characterized in that: The details of feature extraction and large model construction are as follows: Feature extraction: In the context of government big data, identify and extract the most helpful information for a specific task or analysis in the data set, i.e., key features; key features include location information, time series, time type, and subject attributes; Model selection: The selection of the government data model is based on data characteristics and matching accuracy requirements. The characteristics of government data are deeply analyzed, and the required matching accuracy is determined according to business needs and goals. The characteristics of government data include the data structure type, data noise level, and data distribution characteristics. Model training and optimization: Use historical data sets to train the government affairs model, adjust hyperparameters through cross-validation to ensure the generalization ability of the government affairs model; set the initial parameters of the learning rate, batch size, and number of layers of the government affairs model, and select a suitable loss function and optimizer; then based on the performance results of cross-validation, adjust the hyperparameters of the learning rate, batch size, number of layers, and number of hidden units to achieve the best performance, and use an independent test set for final evaluation to ensure that the government affairs model has good generalization ability and can perform well on new data. Closely observe the loss curve and accuracy indicators during the training process of the government affairs model, look for signs of overfitting or underfitting, and take corresponding measures; then based on the feedback from the test set and business needs, continue to adjust the government affairs model, including increasing or decreasing the number of layers, changing the activation function, and adjusting the regularization strategy.

3. According to claim 2, the method for generating intelligent similarity matching of ledgers in urban grassroots governance based on the government affairs big model is characterized in that: Geographic location: Ledger data is usually associated with a specific location, such as the place where the event occurred, the location of the company, and the place where an individual lives; Geographical location features are defined by extracting address information, geographic coordinates or area codes, which are helpful for regional analysis; regional analysis includes urban hotspot distribution and regional resource allocation; Time series: The timestamp information of the extracted occurrence time and recording time is used to analyze the time series characteristics of data change trends, periodic fluctuations and timeliness; Event type: The event type reflects the nature of the ledger entry. By classifying and marking different event types, data can be grouped and compared to find patterns and associations. Subject attributes: Attributes of individuals or organizations related to the ledger. Subject attributes include demographic characteristics, enterprise size, and economic type.

4. According to claim 1, the method for generating intelligent similarity matching of accounts in urban grassroots governance based on the government affairs big model is characterized in that: The specific details of the ledger context-aware matching are as follows: Context information integration, specifically: first, encode the geographic location information into a format understandable to the government big model; the geographic location information includes geographic coordinates and region codes; second, convert the time information of the event into time series data that can be used in the government big model. The time series data includes timestamps and date differences. The time series data is incorporated into the input of the government big model as an additional feature, or introduced at different levels of the government big model to better capture the spatiotemporal correlation behind the event; Condition screening: Screening events based on geographic location, giving priority to events related to specific locations; specifically: when dealing with urban noise issues, give priority to noise complaints near residential areas or schools; at the same time, consider the time factors of working hours and holidays when the events occur, and screen in the time dimension; during holidays, strengthen supervision of the environment around tourist attractions; In addition, incidents are classified and screened based on the type and urgency of the incident to ensure that urgent or high-priority incidents are handled first; Prioritization: Develop context-based rules for event prioritization, specifically: emergency events take precedence over general events, and use learning methods to learn event prioritization patterns based on historical data and automatically adjust priorities; Finally, the severity of the incident, the scope of impact, various attributes of the relevant subjects and contextual information are comprehensively considered to conduct a multi-dimensional evaluation and ranking.

5. According to claim 1, the method for generating intelligent similarity matching of ledgers in urban grassroots governance based on the government affairs big model is characterized in that: Case study and pattern recognition are as follows: Build a comprehensive and detailed case library: systematically collect and organize verified and successfully implemented data matching examples; each case not only includes the final state of successful matching, but also needs to record in detail the key information of input data characteristics, matching strategies adopted and environment configuration during the matching process. Through highly structured data organization, the cases are safely and efficiently stored in advanced database management systems or specially designed knowledge bases, providing a solid foundation for subsequent analysis and learning; Use advanced pattern recognition technology to deeply explore the resources in the case library: involving data governance models, aiming to extract hidden rules and patterns from a large number of successfully matched instances, and through reinforcement learning mechanisms, dynamically adjust the model structure based on the results of pattern recognition; According to the identified effective feature combinations, the feature weights are automatically adjusted to ensure that each match can focus more accurately on the most critical data attributes; based on the identified patterns and rules, the matching government affairs big model is iteratively improved to optimize the structure and parameters of the government affairs big model, and the knowledge obtained from pattern recognition is re-matched.

6. According to claim 1, the method for generating intelligent similarity matching of urban grassroots governance accounts based on the government affairs big model is characterized in that: Ledger intelligent similarity matching also includes real-time updates and learning, as follows: Adaptive weight learning: Design an online learning module to dynamically learn the impact of different features on the matching results during the matching process and automatically adjust the feature weights; And use the newly entered data to continuously optimize the matching algorithm and improve the matching accuracy; Design online learning module: select appropriate online learning algorithm that can process continuously arriving data and gradually update the government model; Determine the appropriate learning rate, control the speed of weight updates, balance the stability of the government model and the need to quickly adapt to new data, and perform dynamic feature weight adjustments, track the impact of different features on matching results, and estimate them through statistical methods or model prediction error analysis; According to the results of feature impact analysis, a weight update mechanism is designed to enhance features that contribute greatly to the matching results and suppress features that contribute little; Introduce regularization terms in weight updates to avoid overfitting caused by excessive weights, maintain the generalization ability of the government affairs model, and continuously optimize matching to include newly entered data in the training set in real time to ensure that the government affairs model can learn from the latest data in real time; Continuously monitor matching accuracy and other relevant performance indicators, promptly detect performance degradation and take action; Regularly retrain the government big model based on accumulated new data to absorb more complex patterns and relationships; Consider using a big government model fusion strategy to combine the judgments of multiple different big government models and improve the robustness of matching through a weighted or voting mechanism.

7. According to claim 1, the method for generating intelligent similarity matching of ledgers in urban grassroots governance based on the government affairs big model is characterized in that: The method also includes an early warning feedback mechanism, as follows: Intelligent early warning and decision support: When the matching results are abnormal or potential problems are found, the early warning is automatically triggered and decision suggestions are provided to the management personnel. The early warning information is promptly notified to the relevant management personnel through the channels of email, SMS or internal message system. At the same time, the learning government affairs big model is used to conduct in-depth analysis of the abnormal causes and provide possible explanations. Performance reports and early warning history records are regularly generated for management personnel to review, and all early warning events and subsequent decision-making operations are recorded to ensure traceability and clear responsibilities. Effect evaluation and feedback loop: Establish an evaluation system for matching results, regularly evaluate the performance of the government affairs big model through the indicators of accuracy and recall rate, collect user feedback, continuously iterate and optimize, and build user-friendly feedback channels such as online surveys, user interviews, and customer service records. Quantify and record user feedback information and analyze it together with the performance indicators of the government affairs big model; conduct in-depth research on specific cases of user feedback; at the same time, determine the priority and direction of optimization based on performance evaluation and user feedback; after the government affairs big model is updated, closely monitor the performance changes and user feedback of the government affairs big model to ensure that the improvement measures achieve the expected results, regularly review the entire evaluation and feedback process, find potential bottlenecks, and further optimize the process.

8. A system for intelligent similarity matching and generation of urban grassroots governance ledgers based on a government affairs big model, characterized by: The system is used to implement the intelligent similarity matching generation method for urban grassroots governance ledgers based on the government affairs big model as described in any one of claims 1 to 7; the system includes: The data integration and preprocessing module is used to collect various raw data related to urban governance through the government cloud platform, and perform preprocessing operations such as cleaning and deduplication on the raw data to obtain the preprocessed raw data; at the same time, encryption technology or desensitization technology is used to process personal or sensitive information to ensure data security and compliance; among them, various raw data related to urban governance include natural person information, real estate registration information, industrial and commercial registration information, urban planning information, and public safety record information; The feature extraction and large model building module is used to extract key features that are highly correlated with the matching degree of the ledger based on the government affairs big data, and use the large language model Transformer to build a matching government affairs big model; The ledger intelligent similarity matching module is used to convert complex contextual information into a format that can be understood by large modules through ledger context-aware matching, and perform precise time screening and priority sorting according to specific conditions to improve governance efficiency and accuracy; then, through case learning and pattern recognition of patterns and rules, the matching large model is iteratively improved, the knowledge obtained from pattern recognition is re-matched for similarity, and updated and learned in real time.

9. An electronic device, characterized in that: include: memory and at least one processor; Wherein, the memory stores computer-executable instructions; The at least one processor executes the computer execution instructions stored in the memory, so that the at least one processor executes the method for generating intelligent similarity matching of ledgers in urban grassroots governance based on the government affairs big model as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer execution instructions. When the processor executes the computer execution, it implements the method for generating intelligent similarity matching of ledgers in urban grassroots governance based on the government affairs big model as described in any one of claims 1 to 7.

Images