Government service recommendation method and system based on user behavior analysis

Abstract

This invention relates to the technical field of government services, and in particular to a method and system for recommending government services based on user behavior analysis. By mining historical and real-time user behavior data, combined with basic user information and policy change information, it achieves personalized and accurate recommendations for government services, improving service efficiency and resource utilization, and enhancing user satisfaction and trust. The method includes: collecting historical user behavior data on a government service platform; performing feature statistics on the historical behavior data to obtain the user's daily behavior baseline for each transaction; acquiring the user's real-time behavior data and extracting features from the real-time behavior data to obtain the user's real-time behavior features for each transaction; in response to any transaction where the user's real-time behavior features exceed the user's daily behavior baseline, extracting all user real-time behavior features exceeding the user's daily behavior baseline, and summarizing them to generate a real-time abnormal behavior feature vector for that user.

Description

Technical Field

[0001] The present invention relates to the technical field of government services, and in particular to a government service recommendation method and system based on user behavior analysis. Background Art

[0002] In the wave of digital transformation, the government service platform serves as an important bridge for interaction between the government and the people. The service quality and efficiency are directly related to public satisfaction and the shaping of the government's image. With the continuous development and popularization of Internet technology, more and more people choose to handle various businesses through government service platforms, which has enabled the government service platforms to accumulate a large amount of user behavior data.

[0003] The existing government service model often focuses on providing standardized service options, mostly staying at the analysis level of a single data source, ignoring the personalized needs and behavioral habits of users, resulting in a lack of targetedness and accuracy in service delivery. At the same time, frequent changes in government service policies also require the platform to be able to respond quickly and adjust service content to adapt to the new policy environment. Summary of the invention

[0004] To solve the above technical problems, the present invention provides a government service recommendation method and system based on user behavior analysis, which mines user history and real-time behavior data, combines user basic information with policy change information, realizes personalized and accurate recommendation of government services, improves service efficiency and resource utilization, and enhances user satisfaction and trust.

[0005] In a first aspect, the present invention provides a method for recommending government services based on user behavior analysis, the method comprising:

[0006] Collect historical behavior data of users on the government service platform;

[0007] Perform feature statistics on historical behavior data to obtain the user's daily behavior baseline corresponding to each transaction;

[0008] Obtain the user's real-time behavior data, extract features from the real-time behavior data, and obtain the user's real-time behavior features corresponding to each transaction;

[0009] In response to a user's real-time behavior feature of any transaction exceeding the user's daily behavior baseline, all user real-time behavior features exceeding the user's daily behavior baseline are extracted and summarized to generate a real-time abnormal behavior feature vector of the user;

[0010] Obtain basic user information and service policy change information, and perform quantitative processing;

[0011] The quantified basic user information and service policy change information are integrated with the real-time abnormal behavior feature vector to obtain a set of government service demand features.

[0012] The government service demand feature set is input into the preset government service recommendation model to obtain the government service items most relevant to the user's needs, and the most relevant government service items are recommended to the user.

[0013] Furthermore, the method for obtaining the user's daily behavior baseline corresponding to each transaction includes:

[0014] Collect historical behavior data of users' daily behavior, and clean the collected historical behavior data to remove abnormal and invalid data;

[0015] Conduct preliminary data analysis on the cleaned historical behavior data to understand the overall distribution and missing data;

[0016] Extract features that reflect user behavior characteristics from historical behavior data, and screen out valuable features through correlation analysis;

[0017] Calculate the mean, median and standard deviation statistics for each feature that reflects the characteristics of user behavior. The mean, median and standard deviation statistics of the features of user behavior characteristics constitute the baseline of user behavior;

[0018] Establish baselines for each service and each user group, including login frequency baseline, browsing depth baseline, action completion rate baseline, and feedback score baseline;

[0019] Set thresholds to identify behaviors that fall outside of the baseline range;

[0020] Use part of the historical data as a training set and the other part as a test set to verify the effectiveness of the user's daily behavior baseline.

[0021] Furthermore, the method for acquiring the real-time behavior characteristics of the user corresponding to each transaction includes:

[0022] Identify users' real-time behavior data and collect user interaction data from government service platforms in real time;

[0023] Use log collectors to process real-time behavioral data generated by the interaction between users and the platform from the government service platform;

[0024] Collect users' real-time behavior data, convert it into a unified format, and perform preliminary filtering and cleaning on the collected real-time data;

[0025] Extract features from the processed real-time user behavior data, including login behavior, browsing behavior, search behavior, service application behavior, interaction behavior, and feedback behavior;

[0026] Based on the extracted behavioral features, the features of the user's real-time behavior are constructed.

[0027] Furthermore, the feature vector of user behavior is as follows:

[0028] ;

[0029] in, Indicates the timestamp of the last login. Indicates the number of logins per day. Indicates the number of pages visited in the current session. Indicates the average time spent on each page. Indicates the page depth visited by the current session. Indicates the number of searches in the current session. Indicates the click-through rate of search results. Indicates the timestamp of the most recent service request. Indicates the number of service requests in the current session. Indicates the number of interactions with the system in the current session. represents the feedback rating provided by the user, Indicates the number of times feedback was provided in the current session. Indicates the search keyword. Indicates the type of service request. Indicates the type of interaction.

[0030] Furthermore, the method for obtaining the government service demand feature set includes:

[0031] Prepare data information on basic user information, service policy change information, and service policy change information;

[0032] All basic user information is quantified, and key elements are extracted from all service policy change information and service policy change information data and converted into numerical values;

[0033] Check whether the data has missing values, correct the abnormal data, and standardize the corrected data;

[0034] The feature selection algorithm is used to select the features related to the government service needs;

[0035] All features are concatenated into a long vector, and different features are given independent weights, and the contribution of the features in the vector is adjusted according to their importance;

[0036] Based on the features obtained through weighted processing, a set of government service demand features is constructed.

[0037] Furthermore, the government service demand characteristics are as follows:

[0038] ;

[0039] in, represents the set of government service demand characteristics, represents the weight of age, represents the weight of gender, Indicates the weight of the profession, represents the weight of service policy changes; represents the weight of real-time abnormal behavior, where , represents the standardized value of age, A quantitative value representing gender, A quantitative value representing an occupation, A standardized value representing a change in service policy, A normalized value that indicates abnormal behavior in real time.

[0040] Furthermore, the method for constructing the government service recommendation model includes:

[0041] Collect users' historical behavior data from the government service platform, process missing historical behavior data and remove invalid data;

[0042] Extract target-related features from historical behavior data;

[0043] Choose the model architecture of the machine learning model based on the problem characteristics and available data;

[0044] The data set is divided into a training set and a test set. The training set data is used to train the selected government service recommendation model, and the test set is used to test the performance of the trained government service recommendation model.

[0045] Adjust the hyperparameters of the government service recommendation model through cross-validation to achieve the best performance;

[0046] Select precision to evaluate the government service recommendation model;

[0047] Collect users' actual feedback on recommended services, and continuously adjust and optimize the government service recommendation model.

[0048] On the other hand, the present application also provides a government service recommendation system based on user behavior analysis, the system comprising:

[0049] The data collection module is used to collect historical and real-time behavior data of users on the government service platform; the collected data types include users' search records, browsing records, click records, and application submission records;

[0050] Feature extraction module: used to perform feature statistics on historical behavior data and calculate the user's daily behavior baseline corresponding to each transaction; analyze the user's latest behavior data in real time and extract the user's real-time behavior features corresponding to each transaction;

[0051] The anomaly detection module is used to compare the real-time behavior features with the daily behavior baseline, identify abnormal behaviors, and summarize the real-time behavior features that exceed the baseline into a real-time abnormal behavior feature vector;

[0052] Information quantification module, used to obtain and quantify the user's basic information and service policy change information, and perform quantitative processing;

[0053] The multi-source data fusion module is used to combine the abnormal behavior feature vector, user basic information and service policy change information to fuse the multi-source data and output the fused government service demand feature set;

[0054] The recommendation model module is used to use the pre-trained government service recommendation model and input the fused feature set to generate a personalized government service recommendation list for the user;

[0055] The recommendation display module is used to display recommended government service items to users; government service items include a feedback mechanism to collect user feedback on the recommended content.

[0056] In a third aspect, the present application provides an electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and the computer program, when executed by the processor, implements the steps of any one of the above methods.

[0057] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps in any one of the above-mentioned methods when executed by a processor.

[0058] Compared with the prior art, the present invention has the following beneficial effects: by collecting and analyzing historical and real-time behavior data of users, it is easier to understand the personalized needs and behavior habits of users; the analysis based on user behavior makes the recommendation of government services more accurate, meets the unique needs of different users, and improves user satisfaction; by considering the frequent changes in government service policies, it can collect and analyze the real-time behavior data of users in real time, and quickly adjust the service recommendation content in combination with policy change information; the dynamic response mechanism enables the government service platform to better adapt to the new policy environment and maintain the timeliness and accuracy of the service; by integrating multi-source data such as basic information of users, real-time behavior data, historical behavior baselines, and service policy change information, it can more comprehensively characterize the needs of users and the service environment, and the comprehensive analysis of multi-source data can improve the recommendation The accuracy and reliability of the system; through the automated recommendation system, the time cost of users manually searching and screening government service items is reduced, and users can quickly obtain the services most relevant to their needs, thereby improving the overall service efficiency of the government service platform; the recommendation method based on user behavior analysis facilitates the government service platform to understand the actual needs and service hotspots of users, and facilitates government agencies to rationally allocate resources, optimize service processes, and improve the overall efficiency of government services; by providing more personalized and accurate government service recommendations, the government service platform can enhance users' trust and satisfaction with government services; by mining user historical and real-time behavior data, combining user basic information with policy change information, the personalized and accurate recommendation of government services can be achieved, service efficiency and resource utilization can be improved, and user satisfaction and trust can be enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is a flow chart of the present invention;

[0060] Figure 2 It is a flow chart of the method for building a government service recommendation model;

[0061] Figure 3 This is the structural diagram of the government service recommendation system based on user behavior analysis. DETAILED DESCRIPTION

[0062] In the description of this application, those skilled in the art should know that this application can be implemented as a method, an apparatus, an electronic device, and a computer-readable storage medium. Therefore, this application can be specifically implemented in the following forms: complete hardware, complete software (including firmware, resident software, microcode, etc.), a combination of hardware and software. In addition, in some embodiments, this application can also be implemented in the form of a computer program product in one or more computer-readable storage media, and the computer-readable storage medium contains computer program code.

[0063] The above-mentioned computer-readable storage medium may adopt any combination of one or more computer-readable storage media. Computer-readable storage media include: electrical, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or devices, or any combination of the above. More specific examples of computer-readable storage media include: portable computer disks, hard disks, random access memories, read-only memories, erasable programmable read-only memories, flash memories, optical fibers, optical disc read-only memories, optical storage devices, magnetic storage devices, or any combination of the above. In the present application, computer-readable storage media can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device.

[0064] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws.

[0065] The present application describes the provided methods, devices, and electronic devices through flowcharts and / or block diagrams.

[0066] It should be understood that each box in the flowchart and / or block diagram and the combination of boxes in the flowchart and / or block diagram can be implemented by computer-readable program instructions. These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine, and these computer-readable program instructions are executed by a computer or other programmable data processing device to produce a device that implements the functions / operations specified by the boxes in the flowchart and / or block diagram.

[0067] These computer-readable program instructions may also be stored in a computer-readable storage medium that enables a computer or other programmable data processing device to work in a specific manner. In this way, the instructions stored in the computer-readable storage medium produce an instruction device product including functions / operations specified in the blocks in the flowchart and / or block diagram.

[0068] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device, so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby enabling the instructions executed on the computer or other programmable data processing apparatus to provide a process for implementing the functions / operations specified in the blocks in the flowchart and / or block diagram.

[0069] The present application is described below in conjunction with the drawings in the present application.

[0070] Embodiment 1: Figure 1 to Figure 2 As shown, the government service recommendation method based on user behavior analysis of the present invention specifically includes the following steps:

[0071] S1. Collect historical behavior data of users on the government service platform;

[0072] Step S1 collects the historical behavior data of users on the government service platform. The collected historical behavior data provides data support for subsequent data analysis and recommendation services; the details are as follows:

[0073] The data of the government service platform comes from the following sources, including:

[0074] User login records, used to record the time, device type, operating system and browser type of each user login; search records, used to record the keywords, search times and search time of users on the government service platform; browsing records: record the pages visited by users, the time spent on each page and the order of visits, etc.; click records, used to record the interactive elements such as links and buttons clicked by users; application submission records, used to record various online applications submitted by users, including application type, submission time and filled content, etc.; feedback records, used to record any feedback provided by users, including evaluations and suggestions, etc.;

[0075] In order to comprehensively and accurately collect historical behavior data of users on the government service platform, the data collection methods adopted include:

[0076] Log files are used to automatically record user behavior data through the log system of the government service platform; front-end monitoring is used to embed JavaScript code or other front-end monitoring tools in web pages or mobile applications to capture user operation behaviors; database query is used to regularly export relevant form data of user submitted information from the database; API interface is used to obtain real-time user behavior data through the API interface

[0077] The collected user behavior data will be reasonably stored and managed for subsequent data analysis and recommendation service use, including removing invalid data, duplicate data and outliers to ensure the accuracy and completeness of the data; classifying and indexing data according to data type and purpose for quick retrieval and query; taking effective measures to protect user privacy and data security to prevent data leakage and abuse; regularly backing up data to prevent loss, and establishing a data recovery mechanism to deal with data loss.

[0078] By collecting data on user login records, search records, browsing records, click records, application submission records, and feedback records, the user's behavioral characteristics on the government service platform are fully reflected, providing rich data support for subsequent data analysis and recommendation services; a combination of log files, front-end monitoring, database queries, and API interfaces is adopted to ensure the accuracy and real-time nature of historical behavior data; different historical behavior data collection methods verify and complement each other to improve the reliability of historical behavior data; in the process of historical behavior data collection, emphasis is placed on the protection of user privacy and data security, data is collected in a legal and compliant manner, and measures are taken to prevent data leakage and abuse, build user trust, and improve the image and reputation of the government service platform; the collected historical behavior data is reasonably stored and managed, including the removal of invalid data, duplicates, and other data. Data and outliers, etc., to ensure the accuracy and completeness of the data, classify and index the data according to the data type and purpose, facilitate rapid retrieval and query of data, and improve the efficiency of data processing; through in-depth analysis of users' historical behavior data, dig out users' personalized needs and preferences, provide a data basis for the government service recommendation system, and be able to recommend relevant services to users more accurately, improve user experience and satisfaction; government service policies change frequently, and through real-time collection and analysis of user behavior data, timely discover the impact of policy changes on user behavior, facilitate the government service platform to respond quickly and adjust service content to adapt to the new policy environment, and improve the flexibility and adaptability of government services; comprehensive data collection methods improve the comprehensiveness, accuracy, privacy protection, efficient management, service personalization and policy adaptability of data collection, and enhance user participation and satisfaction.

[0079] S2. Perform feature statistics on historical behavior data to obtain the user's daily behavior baseline corresponding to each transaction;

[0080] Step S2 performs feature statistics on the collected historical behavior data to obtain the user's daily behavior baseline in various matters; the details are as follows:

[0081] First, the collected historical behavior data is cleaned to remove abnormal and invalid data to ensure the accuracy and completeness of the historical behavior data; the cleaned data is integrated into the database in a unified format for subsequent historical behavior data processing and analysis;

[0082] Conduct preliminary data analysis on historical behavior data to understand the overall distribution and missing data; use charts to display the time series and frequency distribution of user behavior to help understand data characteristics;

[0083] Define indicators that can reflect user behavior characteristics, such as login frequency, page views, service queries, and operation completion rate; extract features that reflect user behavior characteristics from historical behavior data; and screen out valuable features through correlation analysis or principal component analysis;

[0084] Calculate statistics such as mean, median, and standard deviation for each feature, which form the baseline of user behavior; use clustering algorithms to group users to establish specific behavioral baselines for different groups; analyze the changing trends of user behavior over time, identify seasonal patterns and cyclical patterns, etc.

[0085] Establish baselines for each service and each user group, including: login frequency baseline, which indicates the average number of times a user logs into the government service platform in a certain period of time; browsing depth baseline, which indicates the average depth of users browsing specific service pages; operation completion rate baseline, which indicates the success rate of users completing specific service operations; feedback score baseline, which indicates the average score of users on the service;

[0086] Set thresholds to identify behaviors that exceed the baseline range, such as a sudden increase in login frequency or a significant increase in operation failure rate;

[0087] Use part of the historical data as a training set and the other part as a test set to verify the effectiveness of the baseline; have domain experts review the rationality of the baseline to ensure that it conforms to actual business scenarios;

[0088] As user behavior changes and new data accumulates, the behavior baseline is updated regularly; the baseline is adjusted based on user feedback and system performance indicators to adapt to changing needs.

[0089] By removing abnormal and invalid data, the accuracy and completeness of historical behavior data are ensured, providing a reliable data foundation for subsequent analysis; the cleaned data is integrated into a unified database for easy management and analysis; through preliminary data analysis and data visualization, the overall distribution and missingness of the data are understood to help better understand the data characteristics; indicators that reflect the characteristics of user behavior are defined, and features that reflect the characteristics of user behavior are extracted from historical behavior data to prepare for subsequent statistical analysis; statistics such as the mean, median, and standard deviation of each feature are calculated. The above statistics constitute the baseline of user behavior, which is convenient for identifying normal and abnormal behaviors; users are grouped using clustering algorithms, and specific behavior baselines are established for different groups to improve the targetedness of recommendations; user behavior trends over time are analyzed to identify seasonal and cyclical patterns to facilitate the prediction of future behavior patterns; baselines are established for each service and each user group, including login frequency baselines, browsing depth baselines, operation completion rate baselines, and feedback score baselines. The above baselines are used for subsequent anomaly detection and personalized service recommendations. The system provides a reference standard for recommendations; sets thresholds to identify behaviors that exceed the baseline range, such as a sudden increase in login frequency or a significant increase in operation failure rate, so as to facilitate timely detection of abnormal situations; uses part of the historical data as a training set and the other part as a test set to verify the effectiveness of the baseline and ensure the accuracy of the recommendation system; reviews the rationality of the baseline through domain experts to ensure that it conforms to actual business scenarios and improves the practicality and reliability of the user's daily behavior baseline; regularly updates the behavior baseline, adjusts the baseline based on user feedback and system performance indicators, adapts to changing needs, and ensures the continued effectiveness and adaptability of the user's daily behavior baseline; based on the analysis of the user's daily behavior baseline, provides users with more personalized government service recommendations to improve user experience and satisfaction; through the collection of real-time behavior data, quickly responds to changes in user needs and updates to service policies to ensure the timeliness and flexibility of services; Step S2 improves the accuracy of recommendations, enhances system adaptability, improves user experience, and protects data security and privacy, improves the efficiency and quality of government services, and enhances user participation and satisfaction.

[0090] S3. Obtain the user's real-time behavior data, and perform feature extraction on the real-time behavior data to obtain the user's real-time behavior features corresponding to each transaction;

[0091] Step S3 acquires the user's real-time behavior data and extracts features from the real-time behavior data to obtain the user's real-time behavior features corresponding to each transaction, so as to determine whether the user's immediate needs and interest changes can be captured in time; the details are as follows:

[0092] First, identify the user's real-time behavior data. After that, collect the user interaction data from the government service platform in real time, such as clickstream data, search records, form filling behavior, and online customer service conversation records. Use log collectors, API monitoring tools, or real-time data stream processing platforms to capture the data generated by the user's interaction with the platform in real time. Convert the collected real-time data into a unified format for subsequent processing and analysis.

[0093] Perform preliminary filtering and cleaning of the collected real-time data to remove noise data, duplicate data, and obviously erroneous records, and improve the accuracy and efficiency of subsequent feature extraction;

[0094] Feature extraction is performed on the processed real-time behavior data of users, including: login behavior, which is used to record the user's login time and login frequency; browsing behavior, which is used to count the number of pages visited by users, the time they stay on the page and the browsing depth; search behavior, which is used to collect users' search keywords, search frequency and click-through rate of search results; service application behavior, which is used to record the time, application type and application frequency of users submitting service applications; interaction behavior, which is used to record the number and type of interactions between users and the system during the service process; feedback behavior, which is used to collect information such as the number of user feedback and ratings on the service;

[0095] Based on the real-time user behavior features collected above, construct features that reflect the user's current behavior characteristics; use statistical methods, machine learning methods or domain knowledge to filter out features related to user behavior; use one-hot encoding or label encoding to encode non-numerical features;

[0096] Use real-time data analysis tools to analyze the collected behavior data in real time to quickly identify users' immediate needs; update the feature vector of user behavior in real time based on the latest user behavior data to ensure that the input is the latest user status;

[0097] The feature vector of user behavior is as follows:

[0098] ;

[0099] in, The feature vector representing the user behavior, Indicates the timestamp of the last login. Indicates the number of logins per day. Indicates the number of pages visited in the current session. Indicates the average time spent on each page. Indicates the page depth visited by the current session. Indicates the number of searches in the current session. Indicates the click-through rate of search results. Indicates the timestamp of the most recent service request. Indicates the number of service requests in the current session. Indicates the number of interactions with the system in the current session. represents the feedback rating provided by the user, Indicates the number of times feedback was provided in the current session. Indicates the search keyword. Indicates the type of service request. Indicates the type of interaction.

[0100] Through tools such as real-time data stream processing platforms and log collectors, users' real-time behaviors are captured to ensure timely response to changes in user needs; real-time data analysis tools are used to quickly identify users' real-time needs and improve the response speed of the recommendation system; through preliminary data filtering and cleaning, noise data, duplicate data and obviously erroneous records are removed to improve the accuracy and efficiency of subsequent feature extraction; the collected real-time data is converted into a unified format to facilitate subsequent processing and analysis; feature extraction is performed on the processed real-time behavior data, covering login behavior, browsing behavior, search behavior, service application behavior, interactive behavior and feedback behavior, to ensure the comprehensiveness and representativeness of the features; statistical methods, machine learning methods or domain knowledge are used to screen out features that are highly correlated with user behavior and improve the quality of feature vectors ; Encode non-numeric features so that the features can be effectively processed by the machine learning model; Update the user's feature vector in real time according to the latest user behavior data to ensure that the recommendation system always makes recommendations based on the latest user status; By capturing the user's immediate needs and interest changes in real time, quickly respond to changes in user behavior and updates to service policies, and provide users with more personalized government service recommendations, ensure the timeliness and flexibility of services and improve user experience and satisfaction; Through efficient data cleaning and conversion technology, ensure the quality and availability of data; By quickly adjusting service content according to changes in service policies, ensure consistency with the latest policy environment; Step S3 improves the advantages in solving the problem of capturing immediate needs based on real-time user behavior analysis in the government service platform, and improves the efficiency and quality of government services.

[0101] S4, in response to any transaction, the user's real-time behavior feature exceeds the user's daily behavior baseline, extracting all the user's real-time behavior features that exceed the user's daily behavior baseline, and summarizing and generating a real-time abnormal behavior feature vector of the user;

[0102] Step S4 is a process of identifying the difference between the user's real-time behavior and their daily behavior and extracting abnormal behavior features from them; the details are as follows:

[0103] In step S2, the user's behavior baseline in daily operation is obtained by statistically analyzing the historical behavior data. The user's daily behavior baseline represents the user's behavior pattern and habits under normal circumstances;

[0104] Collect the real-time behavior data of the user obtained in step S3, and compare the real-time behavior data of the user with the daily behavior baseline of the user obtained in step S2; use statistical tests to determine whether there is a significant difference between the real-time behavior data and the daily behavior baseline;

[0105] Through historical data analysis or expert experience and based on actual scenarios and data distribution, a reasonable threshold is set to determine whether the user's real-time behavior exceeds the user's daily behavior baseline based on the set threshold;

[0106] When the user's real-time behavior data exceeds the threshold range, it is considered that the feature exhibits abnormal behavior; all real-time behavior features that exceed the daily behavior baseline threshold are extracted. When extracting abnormal behavior features, the accuracy and completeness of the information are ensured for subsequent analysis and processing;

[0107] The real-time behavior data characteristics of the user are: login behavior, including abnormal login time or frequency; browsing behavior, including abnormal page visits, browsing depth or dwell time; search behavior, including abnormal search keywords, frequency or click-through rate; service application behavior, including abnormal service application time or type; interaction behavior, including abnormal interaction times or types; feedback behavior, including abnormal feedback or ratings;

[0108] All the extracted abnormal behavior features are aggregated together and combined into the user's real-time abnormal behavior feature vector in order or according to rules; in the process of combining into the real-time abnormal behavior feature vector, the weight relationship between different features is considered to ensure that the vector can truly reflect the user's abnormal behavior.

[0109] By comparing the user's real-time behavior with the daily behavior baseline, accurately identify which behavioral characteristics show abnormalities; use statistical tests to determine whether there are significant differences between real-time behavior data and daily behavior baselines to improve the accuracy of anomaly detection; by real-time monitoring of users' online behavior, abnormal behavior can be discovered in a timely manner. Real-time monitoring makes it easier to respond to potential problems or threats more quickly; set reasonable thresholds based on historical data analysis or expert experience to adapt to different application scenarios and needs. The threshold setting is adjusted according to the actual scenario and data distribution to improve flexibility; consider various types of user behavior characteristics, including login, browsing, search, service application, interaction and feedback, etc., to comprehensively capture user behavior. The method can monitor the online activities of users; comprehensively evaluate the changes in user behavior by monitoring these diverse features; extract and summarize abnormal behavior features to form a structured real-time abnormal behavior feature vector to facilitate further analysis and processing. The structured vector form is helpful for subsequent automated processing and algorithm application; by extracting abnormal behavior features and forming vectors, clear information is provided to help them quickly understand the problem; the abnormal behavior feature vector provided helps security teams or data analysts make response decisions faster; step S4 is used to identify the difference between the user's real-time behavior and daily behavior, and extract abnormal behavior features therefrom, which is accurate and flexible in identifying abnormal behavior features in the user's real-time behavior.

[0110] S5. Obtain basic user information and service policy change information, and perform quantitative processing;

[0111] Step S5 is the process of obtaining basic user information and service policy change information and performing quantitative processing; the details are as follows:

[0112] The process of obtaining and quantifying basic user information includes:

[0113] The user's basic information comes from registration information or data obtained from other credible channels; the user's basic information includes: user identity attributes, such as age, gender, occupation and place of residence, which help to understand the general characteristics and demand differences of the user group; historical service records: the type, frequency and satisfaction of the user's past use of government services, which reflect the user's interests and potential needs; preference information that the user actively sets on the platform, such as language preference and notification method, which reflects the user's personalized needs;

[0114] In order to make the user's basic information understandable and processable by computers, it needs to be converted into numerical form; quantified in the following ways: categorical variables are converted using one-hot encoding or label encoding; continuous variables are converted into categorical variables using original values ​​or after binning; geographic locations are represented by latitude and longitude coordinates, or processed according to administrative division codes;

[0115] The process of obtaining and quantifying information on service policy changes includes:

[0116] Government service policies are constantly adjusted and improved by the government according to the needs of social development. Changes in service policies directly affect the content and methods of government services. Ways to obtain information on service policy changes include: policy release platforms, regularly capturing the latest policy document news reports from government official websites, policy release platforms and other channels, and timely obtaining policy change information through news reports; third-party data sources, using policy change information provided by professional policy research institutions or data providers;

[0117] The process of quantifying service policy change information includes: extracting key elements from the policy text, such as the effective date of the new policy, the applicable population, and the specific service items involved; constructing policy features to convert the extracted key elements into numerical or categorical features, such as using binary variables to indicate whether a policy is applicable to a specific group, or using numerical values ​​to indicate the urgency of the policy; time series processing, when policy changes are time-dependent, time series processing is performed by constructing time series features, such as the frequency of policy updates, the time interval between the most recent policy update and the current update, etc.

[0118] After completing the quantitative processing of user basic information and service policy change information, these two parts of data need to be integrated to better carry out subsequent analysis and modeling; processing the quantified data involves preprocessing steps such as data cleaning, missing value processing, and outlier detection.

[0119] By using the user's identity attributes to accurately identify the differences in needs of different user groups, more personalized service recommendations can be provided; historical service records make it easier to understand user preferences and service habits, and to predict future service needs; the use of user preference information can improve user experience and make users feel humane service; by using quantitative methods such as unique hot encoding and label encoding, categorical variables can be converted into machine-readable forms to facilitate subsequent analysis and modeling; by quantifying geographic location information, it is easier to provide geographic location-related service recommendations, such as nearby service outlets, local policy information, etc.; regularly grabbing policy information from multiple channels ensures the timeliness of policy updates, allowing the platform to quickly adapt to the new policy environment and reduce the inconvenience caused by policy lags; by extracting key information The policy characteristics are constructed to accurately understand the policy intentions and scope of influence, so as to better serve users; the policy changes are quantified and integrated into user characteristics to ensure that the service items recommended to users are user-related information; time series processing facilitates the identification of trends and periodicity of policy changes, thereby predicting future policy directions and changes in service needs; integrating user basic information and service policy change information can create a more comprehensive user portrait, which is convenient for providing accurate service recommendations; step S5 can improve the personalization level of government services, enhance user experience, speed up the response to policy changes, and improve the relevance and accuracy of service recommendations; through optimization measures, the government service platform can better meet user needs and improve the overall quality and efficiency of public services.

[0120] S6. Perform multi-source data fusion on the quantified user basic information and service policy change information and the real-time abnormal behavior feature vector to obtain a set of government service demand features;

[0121] Step S6 is to combine the quantified user basic information and service policy change information with the real-time abnormal behavior feature vector to form a government service demand feature set; the details are as follows:

[0122] S61. Prepare data information of user basic information, service policy change information and service policy change information; ensure that all user basic information has been quantified, such as age, gender and occupation; ensure that all service policy change information has been extracted with key elements and converted into numerical or categorical features; ensure that abnormal behavior feature vectors have been extracted and contain all behavior features that exceed the daily behavior baseline;

[0123] S62. Check whether there are missing values ​​in the above data, and perform appropriate filling or deletion processing; remove or correct abnormal data; standardize or normalize the above data to ensure comparability between different features;

[0124] S63. Use feature selection algorithms to select the features most relevant to government service needs; if there are too many features, use dimensionality reduction techniques such as principal component analysis or autoencoders to reduce the number of features and try to keep important information;

[0125] S64. All features are directly concatenated together to form a long vector. Different weights are assigned to different features, and the contribution in the vector is adjusted according to the importance of the features. The correlation between different features is captured by cross-features, such as the interaction between age and occupation affecting the preference for certain services.

[0126] S65. Construct a final government service demand feature set based on the features obtained through the above processing; the feature set includes all processed basic user information, service policy change information, and real-time abnormal behavior features;

[0127] The government service demand characteristics are as follows:

[0128] ;

[0129] in, represents the set of government service demand characteristics, represents the weight of age, represents the weight of gender, Indicates the weight of the profession, represents the weight of service policy changes; represents the weight of real-time abnormal behavior, where , represents the standardized value of age, A quantitative value representing gender, A quantitative value representing an occupation, A standardized value representing a change in service policy, Normalized values ​​that indicate abnormal behavior in real time;

[0130] S66. Before actual application, use cross-validation method to evaluate the quality of feature set.

[0131] By combining basic user information, service policy change information and real-time abnormal behavior characteristics, it is possible to fully capture the multi-dimensional needs of users, provide data support for government service recommendations, and facilitate the improvement of the accuracy and personalization of recommendations; through quantitative processing and feature selection, it is possible to remove redundant and noisy data and retain the features most relevant to government service needs; by assigning different weights to different features and capturing the correlation between features, the accuracy of recommendations is further improved; by quickly responding to changes in service policies and real-time changes in user behavior, updating abnormal behavior feature vectors and policy change information, it ensures that the recommendation system is always consistent with the latest service environment and user needs; through steps such as cross-validation and feature selection, it is possible to provide highly interpretable recommendation results; it helps to improve users' trust in the recommendation system and government agencies understand the changing trends of user needs, and provide reference for policy formulation and service optimization; the data preprocessing stage emphasizes the importance of data quality, and improves the efficiency and accuracy of subsequent data analysis by checking missing values, processing abnormal data and performing data standardization; through the cross-validation evaluation mechanism, it is convenient to evaluate the quality of the feature set and the performance of the recommendation system, evaluate the effects of different feature combinations and model parameters, select the optimal recommendation plan, and further improve user satisfaction and government service efficiency.

[0132] S7, inputting the government service demand feature set into a preset government service recommendation model, obtaining the government service items most relevant to the user's needs, and recommending the most relevant government service items to the user;

[0133] Step S7 predicts the government service items that the user is most likely to need at the moment through a preset government service recommendation model based on the government service demand feature set generated in step S6, and recommends them to the user accordingly, as follows:

[0134] First, we build a government service recommendation model, including:

[0135] Collect historical behavior data of users from the government service platform, such as user browsing history, search history, and service request records; remove invalid data, handle missing values, and detect outliers; extract useful features from historical behavior data, such as user preferences, service popularity, and time series characteristics of user behavior;

[0136] Select features relevant to the target and reduce irrelevant or redundant features; construct new derived features based on existing features to enhance the expressiveness of the model; encode categorical features, such as one-hot encoding or label encoding;

[0137] Choose the appropriate model architecture based on the problem characteristics and available data. The model architecture includes traditional machine learning models and deep learning models.

[0138] The data set is divided into a training set and a test set. The training set data is used to train the selected government service recommendation model, and the test set is used to test the performance of the trained government service recommendation model. The hyperparameters of the government service recommendation model are adjusted through cross-validation to obtain the best performance.

[0139] Choose appropriate evaluation indicators to evaluate the government service recommendation model, such as precision, recall, F1 score, average precision or root mean square error. Before deploying the model in the actual environment, you can use A / B testing to compare the effects of the new and old models.

[0140] Design the architecture for the deployment of the government service recommendation model to ensure that the government service recommendation model can quickly respond to user requests; implement real-time prediction capabilities to provide instant recommendations when users access the government service platform; continuously monitor the performance of the government service recommendation model after deployment, and make necessary adjustments and maintenance based on feedback;

[0141] Collect actual feedback from users on recommended services, and continuously adjust and optimize the government service recommendation model based on user feedback to improve the quality of recommendations;

[0142] Then, the government service demand feature set obtained in step S6 is input into the government service recommendation model in the form of a vector;

[0143] After receiving the set of government service demand features, the government service recommendation model uses the internally learned patterns and rules to predict the government services that users may be interested in.

[0144] The predicted government service items are presented to the user through the front-end interface of the government service platform in a ranking list, and the ranking list lists a number of government service items that are most relevant to the user's current needs;

[0145] Collect user feedback on recommended government services, including indicators such as click-through rate and completion rate; continuously adjust and optimize the government service recommendation model based on user feedback to improve the accuracy and satisfaction of recommendations.

[0146] By collecting and analyzing users' historical behavior data, building user portraits, and combining real-time behavior data with service policy changes, personalized government service recommendations can be achieved, which can more accurately capture users' real needs and improve the pertinence and accuracy of recommendations; using advanced machine learning or deep learning models, it can quickly process and analyze large amounts of user data, respond to user requests in real time, provide instant recommendation services, improve the overall service efficiency of the government service platform, and reduce user waiting time; the government service recommendation model can be dynamically adjusted according to frequent changes in service policies to ensure that the recommended content is always consistent with the latest policies, so that it can adapt to the complex and changing government environment and maintain the timeliness and accuracy of recommended services; by collecting user feedback on recommended services, constantly adjusting and optimizing the recommendation model, it helps to improve the quality of recommendations, approach users' real needs, and improve user satisfaction and loyalty; the design of the government service recommendation model takes into account Consider possible future data growth and changes in business needs, support the expansion of new features and new data, and improve competitiveness and innovation in future development; in the process of building and deploying the government service recommendation model, use a variety of evaluation indicators and methods to comprehensively evaluate the performance of the government service recommendation model, which helps to ensure that the government service recommendation model performs well in accuracy, recall rate, F1 score and other aspects, thereby providing users with high-quality recommendation services; by presenting the predicted government service items to users in the form of a ranked list, users can intuitively see the service items most relevant to their current needs, improve user experience, and reduce the difficulty of users screening and finding services; step S7 has the advantages of personalization, efficiency, flexibility, continuous optimization, scalability, comprehensive performance evaluation and user experience optimization when solving the government service recommendation problem in the government service platform, thereby improving public satisfaction with the level of government services.

[0147] Embodiment 2: Figure 3 As shown, the government service recommendation system based on user behavior analysis of the present invention specifically includes the following modules:

[0148] The data collection module is used to collect historical and real-time behavior data of users on the government service platform; the collected data types include users' search records, browsing records, click records, and application submission records;

[0149] Feature extraction module: used to perform feature statistics on historical behavior data and calculate the user's daily behavior baseline corresponding to each transaction; analyze the user's latest behavior data in real time and extract the user's real-time behavior features corresponding to each transaction;

[0150] The anomaly detection module is used to compare the real-time behavior features with the daily behavior baseline, identify abnormal behaviors, and summarize the real-time behavior features that exceed the baseline into a real-time abnormal behavior feature vector;

[0151] Information quantification module, used to obtain and quantify the user's basic information and service policy change information, and perform quantitative processing;

[0152] The multi-source data fusion module is used to combine the abnormal behavior feature vector, user basic information and service policy change information to fuse the multi-source data and output the fused government service demand feature set;

[0153] The recommendation model module is used to use the pre-trained government service recommendation model and input the fused feature set to generate a personalized government service recommendation list for the user;

[0154] The recommendation display module is used to display recommended government service items to users; government service items include a feedback mechanism to collect user feedback on the recommended content.

[0155] By analyzing the personalized needs and behavioral habits of users, we provide service recommendations that meet their actual needs, thereby improving user satisfaction and the efficiency of government service utilization; by considering the real-time behavioral data of users, combined with the changes in basic user information and service policies, we achieve multi-dimensional data analysis and improve the accuracy and relevance of recommendations; by quickly responding to policy changes and adjusting recommended content in a timely manner, we ensure that recommended service items are consistent with the latest policies and help users adapt to new policies more quickly; the anomaly detection module identifies anomalies in user behavior, facilitates the discovery of potential needs or problems that users may encounter, and improves service quality; dynamically adjusts recommendation strategies and service content based on user feedback to form a closed loop of continuous improvement; Personalized recommendations can reduce the time users spend sifting through numerous service options, thereby improving service efficiency and overall user experience. Efficient service push can enhance the public’s trust in and satisfaction with government agencies, and facilitate the shaping of a positive government image. Through in-depth analysis of user behavior data, we can better understand the distribution of user needs, rationally allocate resources, optimize service processes, and reduce unnecessary waste of resources. Based on the results of data analysis, we can make more scientific and reasonable decisions to meet the needs of more users. The government service recommendation system based on user behavior analysis can effectively solve the problems existing in the existing government service model, and through the use of big data and artificial intelligence technology, it can realize the intelligence and personalization of government services and improve the quality and efficiency of government services.

[0156] The various variations and specific embodiments of the government service recommendation method based on user behavior analysis in the aforementioned embodiment 1 are also applicable to the government service recommendation system based on user behavior analysis in this embodiment. Through the aforementioned detailed description of the government service recommendation method based on user behavior analysis, those skilled in the art can clearly understand the implementation method of the government service recommendation system based on user behavior analysis in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here.

[0157] In addition, the present application also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are respectively connected via a bus. When the computer program is executed by the processor, each process of the above-mentioned method for controlling output data is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0158] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A government service recommendation method based on user behavior analysis, characterized in that: The method comprises: Collect historical behavior data of users on the government service platform; Perform feature statistics on historical behavior data to obtain the user's daily behavior baseline corresponding to each transaction; Obtain the user's real-time behavior data, extract features from the real-time behavior data, and obtain the user's real-time behavior features corresponding to each transaction; In response to a user's real-time behavior feature of any transaction exceeding the user's daily behavior baseline, all user real-time behavior features exceeding the user's daily behavior baseline are extracted and summarized to generate a real-time abnormal behavior feature vector of the user; Obtain basic user information and service policy change information, and perform quantitative processing; The quantified basic user information and service policy change information are integrated with the real-time abnormal behavior feature vector to obtain a set of government service demand features. The government service demand feature set is input into the preset government service recommendation model to obtain the government service items most relevant to the user's needs, and the most relevant government service items are recommended to the user.

2. The method for recommending government services based on user behavior analysis according to claim 1, characterized in that: The method for obtaining the user's daily behavior baseline corresponding to each transaction includes: Collect historical behavior data of users' daily behavior, and clean the collected historical behavior data to remove abnormal and invalid data; Conduct preliminary data analysis on the cleaned historical behavior data to understand the overall distribution and missing data; Extract features that reflect user behavior characteristics from historical behavior data, and screen out valuable features through correlation analysis; Calculate the mean, median and standard deviation statistics for each feature that reflects the characteristics of user behavior. The mean, median and standard deviation statistics of the features of user behavior characteristics constitute the baseline of user behavior; Establish baselines for each service and each user group, including login frequency baseline, browsing depth baseline, action completion rate baseline, and feedback score baseline; Set thresholds to identify behaviors that fall outside of the baseline range; Use part of the historical data as a training set and the other part as a test set to verify the effectiveness of the user's daily behavior baseline.

3. The method for recommending government services based on user behavior analysis according to claim 1, characterized in that: The method for acquiring the real-time behavior characteristics of the user corresponding to each transaction includes: Identify users' real-time behavior data and collect user interaction data from government service platforms in real time; Use log collectors to process real-time behavioral data generated by the interaction between users and the platform from the government service platform; Collect users' real-time behavior data, convert it into a unified format, and perform preliminary filtering and cleaning on the collected real-time data; Extract features from the processed real-time user behavior data, including login behavior, browsing behavior, search behavior, service application behavior, interaction behavior, and feedback behavior; Based on the extracted behavioral features, the features of the user's real-time behavior are constructed.

4. The method for recommending government services based on user behavior analysis according to claim 3, characterized in that: The feature vector of user behavior is as follows: ； in, Indicates the timestamp of the last login. Indicates the number of logins per day. Indicates the number of pages visited in the current session. Indicates the average time spent on each page. Indicates the page depth visited by the current session. Indicates the number of searches in the current session. Indicates the click-through rate of search results. Indicates the timestamp of the most recent service request. Indicates the number of service requests in the current session. Indicates the number of interactions with the system in the current session. represents the feedback rating provided by the user, Indicates the number of times feedback was provided in the current session. Indicates the search keyword. Indicates the type of service request. Indicates the type of interaction.

5. The method for recommending government services based on user behavior analysis according to claim 1, characterized in that: The method for obtaining the government service demand feature set includes: Prepare data information on basic user information, service policy change information, and service policy change information; All basic user information is quantified, and key elements are extracted from all service policy change information and service policy change information data and converted into numerical values; Check whether the data has missing values, correct the abnormal data, and standardize the corrected data; The feature selection algorithm is used to select the features related to the government service needs; All features are concatenated into a long vector, and different features are given independent weights, and the contribution of the features in the vector is adjusted according to their importance; Based on the features obtained through weighted processing, a set of government service demand features is constructed.

6. The method for recommending government services based on user behavior analysis according to claim 5, characterized in that: The government service demand characteristics are as follows: ； in, represents the set of government service demand characteristics, represents the weight of age, represents the weight of gender, Indicates the weight of the profession, represents the weight of service policy changes; represents the weight of real-time abnormal behavior, where , represents the standardized value of age, A quantitative value representing gender, A quantitative value representing an occupation, A standardized value representing a change in service policy, A normalized value that indicates abnormal behavior in real time.

7. The method for recommending government services based on user behavior analysis according to claim 1, characterized in that: The method for constructing the government service recommendation model includes: Collect users' historical behavior data from the government service platform, process missing historical behavior data and remove invalid data; Extract target-related features from historical behavior data; Choose the model architecture of the machine learning model based on the problem characteristics and available data; The data set is divided into a training set and a test set. The training set data is used to train the selected government service recommendation model, and the test set is used to test the performance of the trained government service recommendation model. Adjust the hyperparameters of the government service recommendation model through cross-validation to achieve the best performance; Select precision to evaluate the government service recommendation model; Collect users' actual feedback on recommended services, and continuously adjust and optimize the government service recommendation model.

8. A government service recommendation system based on user behavior analysis, characterized in that: The system comprises: The data collection module is used to collect historical and real-time behavior data of users on the government service platform; the collected data types include users' search history, browsing history, click history, and application submission history; Feature extraction module: used to perform feature statistics on historical behavior data and calculate the user's daily behavior baseline corresponding to each transaction; analyze the user's latest behavior data in real time and extract the user's real-time behavior features corresponding to each transaction; The anomaly detection module is used to compare the real-time behavior features with the daily behavior baseline, identify abnormal behaviors, and summarize the real-time behavior features that exceed the baseline into a real-time abnormal behavior feature vector; Information quantification module, used to obtain and quantify users' basic information and service policy change information; The multi-source data fusion module is used to combine the real-time abnormal behavior feature vector, the quantified user basic information and the service policy change information, perform multi-source data fusion, and output the fused government service demand feature set; The recommendation model module is used to use the pre-trained government service recommendation model and input the fused feature set to generate a personalized government service recommendation list for the user; The recommendation display module is used to display recommended government service items to users; government service items include a feedback mechanism to collect user feedback on the recommended content.

9. An electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, wherein: When the computer program is executed by the processor, the steps in the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method according to any one of claims 1 to 7 are implemented.

Images