Government service intelligent recommendation method, system, device and storage medium

Through the intelligent recommendation method of government service, multi-dimensional data and feature computing technology are used to provide users with personalized recommendations, solving the problem of inaccurate recommendations of existing government service websites and improving the accuracy of user experience and recommendations.

CN116975396BActive Publication Date: 2024-12-03BEIJING BIG DATA CENT
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311237996.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-09-25
Publication Date
2024-12-03
Estimated Expiration
2043-09-25

AI Technical Summary

Technical Problem

Existing government service websites cannot provide users with personalized recommendations, which leads to users needing to manually search for a large amount of information when querying and handling business, which consumes time and energy, and the information recommendation is inaccurate, and important content may be missed.

Method used

Using the intelligent recommendation method of government services, the user characteristics are obtained by collecting multi-dimensional data of the target user, pre-processing, and the preference feature vector is calculated based on the average method and the time attenuation method, attribute features of the content to be recommended are extracted, and the content similarity is calculated based on the user's preference features. , generate a recommendation list to realize personalized recommendations.

Benefits of technology

It improves the accuracy of the matching degree between users and the content to be recommended, enhances the accuracy of recommendations, provides a personalized service experience, and meets user needs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116975396B_ABST
    Figure CN116975396B_ABST
Patent Text Reader

Abstract

The present application relates to a method, system, device and storage medium for intelligent recommendation of government services, and relates to the fields of smart cities and e-government, wherein the method comprises: collecting multi-dimensional data of target users, and preprocessing the data to obtain user characteristics of the target users; calculating the preference feature vector of the target users by the time decay method according to the user characteristics of the target users; extracting attribute characteristics of the content to be recommended, and calculating the similarity between the content to be recommended in combination with the preference feature vector of the target user; calculating the matching degree between the target user and each content to be recommended based on the similarity between the preference feature vector of the target user and the content to be recommended, and generating a recommendation list; the problem that personalized services cannot be provided to users and content recommendations are inaccurate in the field of government services can be solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the fields of smart cities and e-government, and in particular to a method, system, device and storage medium for intelligent recommendation of government services. Background Art

[0002] In the field of government services, the existing government service network covers a large amount of user data and business processing information. However, at this stage, government service data is disordered and fragmented, and the problem of data silos is still prominent, which has hindered the advancement of government service work to a certain extent.

[0003] At present, when users search for the services they need to handle on government service websites, they often need to manually search and browse a large amount of information and materials. The whole process is very time-consuming and labor-intensive. In addition, due to the huge amount of information on the service website, users may not be able to determine the service that best suits their needs during the search process, and may also miss important information. Although some cities have opened intelligent recommendation sections, many of them are based on the latest services opened for users, which cannot truly meet the needs of users and fail to provide personalized services to users.

[0004] Therefore, how to provide personalized services to users to improve their satisfaction and experience is an urgent problem to be solved. Summary of the invention

[0005] The purpose of this application is to provide a method, system, device and storage medium for intelligent recommendation of government services to solve the problem that personalized services cannot be provided to users and content recommendations are inaccurate in the field of government services.

[0006] In a first aspect, an embodiment of the present invention provides a method for intelligently recommending government services, the method comprising:

[0007] Collecting multi-dimensional data of target users, and preprocessing the data to obtain user characteristics of the target users;

[0008] According to the user characteristics of the target user, a preference feature vector of the target user is calculated based on an averaging method and a time decay method;

[0009] Extracting attribute features of the content to be recommended, and calculating the similarity between the features of the content to be recommended in combination with the preference feature vector of the target user;

[0010] Based on the preference feature vector of the target user and the similarity between the features of the content to be recommended, the matching degree between the target user and each content to be recommended is calculated to generate a recommendation list.

[0011] Optionally, the calculating the preference feature vector of the target user based on the user characteristics of the target user and the averaging method and the time decay method includes:

[0012] Obtain historical behavior data of target users and transform the historical behavior data into multi-dimensional feature vectors;

[0013] Calculate the time decay weight of the feature vector corresponding to the historical behavior data at each time point based on the multidimensional feature vector;

[0014] The product of the time decay weight and the feature vector at the corresponding time point is calculated based on the averaging method, and the product result is used as the preference feature vector of the target user.

[0015] Optionally, the acquiring of historical behavior data of the target user and transforming the historical behavior data into a multi-dimensional feature vector includes:

[0016] Obtaining the operation record of the target user and the time point of performing the operation recorded in the historical behavior data;

[0017] Based on the operation records and the time point of executing the operation, the historical behavior data of the target user is constructed into a multi-dimensional feature vector matrix; the dimension of the feature vector matrix is ​​determined based on the number of the operation records.

[0018] Optionally, the product result of calculating the time attenuation weight and the feature vector of the corresponding time point based on the averaging method can be expressed as: N represents the total number of contents in the recorded user's historical behaviors; The feature vector representing the i-th content that the user likes; α is the attenuation factor, which takes a value less than 1. is the time interval from when the user takes action on the i-th content to the current time. Optionally, extracting the attribute features of the content to be recommended and calculating the similarity between the features of the content to be recommended in combination with the preference feature vector of the target user includes:

[0019] Extracting attribute features of the content to be recommended and generating a target feature vector;

[0020] Obtaining a preference feature vector of the target user;

[0021] The similarity between the target feature vectors and the similarity between the preference feature vectors are calculated respectively, and the similarity between the features of the content to be recommended is obtained by summing them up.

[0022] Optionally, the calculation of the matching degree between the target user and each content to be recommended can be expressed as: in, represents the target user's preference feature vector, is the attribute feature vector of the content to be recommended, k represents the kth feature, and there are K features in total. Optionally, before collecting the multi-dimensional data of the target user, the user features of the target user are managed, including: pre-defining the user's feature attributes, and analyzing and storing the data according to the preset feature attributes.

[0023] In a second aspect, the present invention provides a government service intelligent recommendation system, the system comprising:

[0024] A user feature acquisition module, used to collect multi-dimensional data of target users and pre-process the data to obtain user features of the target users;

[0025] A preference feature vector generation module, used to calculate the preference feature vector of the target user based on the user characteristics of the target user and the averaging method and the time decay method;

[0026] A similarity calculation module is used to extract the attribute features of the content to be recommended, and calculate the similarity between the features of the content to be recommended in combination with the preference feature vector of the target user;

[0027] The recommendation list generation module is used to calculate the matching degree between the target user and each content to be recommended based on the preference feature vector of the target user and the similarity between the features of the content to be recommended, and generate a recommendation list.

[0028] In a third aspect, the present invention provides a recommendation device, the device comprising:

[0029] one or more processors;

[0030] The memory is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement any of the government service intelligent recommendation methods described above.

[0031] In a fourth aspect, an embodiment of the present invention further provides a storage medium comprising computer executable instructions, which, when executed by a computer processor, are used to execute any of the government service intelligent recommendation methods described above.

[0032] The beneficial effects of the present invention are as follows: the embodiments of the present invention determine the recommended content that the target user may be interested in based on the user characteristics of the target user, calculate the matching degree between the target user and each content to be recommended, and then determine the target content that can be recommended to the target user. This can provide the user with personalized recommendation services based on the user's needs, preferences and historical behaviors, improve the accuracy of the matching degree calculation between the user and the content to be recommended, and then improve the accuracy of the content recommended to the user.

[0033] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A flowchart of a method for intelligently recommending government services provided by an embodiment of the present invention;

[0035] Figure 2 A schematic diagram of the overall framework of a method for intelligently recommending government services provided by an embodiment of the present invention;

[0036] Figure 3 A schematic diagram of the structure of an intelligent government service recommendation system provided by an embodiment of the present invention;

[0037] Figure 4 A schematic diagram of the structure of a recommended device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only the parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0039] The present invention aims at the field of government services. The existing government service network is unable to provide personalized recommendation services to users. Therefore, a government service intelligent recommendation method is proposed. The method is applied to the government service applet, which can provide users with personalized recommendation services in real time to improve user satisfaction and experience.

[0040] Figure 1 This is a flowchart of a method for intelligently recommending government services provided by an embodiment of the present invention. This embodiment is applicable to recommending matters that may be required to be handled by users. The method can be executed by a recommendation system, which can be configured in a recommendation device. Figure 1 As shown, the method includes:

[0041] Step S101: collect multi-dimensional data of target users, and pre-process the data to obtain user characteristics of the target users.

[0042] It should be noted that the target user is the user who operates the government service mini program.

[0043] The multi-dimensional data of the target user, including the user's basic data, user behavior data, and user portrait data, the data source is connected to the external data source, the collected user data is saved in the database, and the collected data is structured and managed. Among them, the basic user data may include the user's age, place of residence, certificate and work information, etc. User behavior data can represent data related to user behavior. For example, browsing the policy query page, applying for project declaration, searching for certificates and commenting on notifications on the management of high-tech enterprise certification can all be regarded as user behavior. Behavioral data is the data generated by users when performing these user behaviors. Exemplarily, the database can be a starrocks database based on the MPP architecture, which analyzes and stores user data.

[0044] Specifically, when storing user data in a database, a user identifier is set for each user, and the user identifier is used to uniquely identify the user so as to distinguish and locate the user at a single point.

[0045] After the collected user data is stored in the database, the data is cleaned through the polling service, mainly including removing duplicate data, processing missing values, processing outliers, etc. to ensure the quality and consistency of the data. At the same time, in order to extract user features later, the data needs to be pre-processed by normalization, standardization, dimensionality reduction, etc.

[0046] The characteristics of the target user are used to characterize the user's interests, preferences, and needs for content. The content can be matters handled or browsed by the user. For example, matters can be policy inquiries, project applications, and certificate application pages. After the user's data is preprocessed, feature engineering technology is used to extract the user's attribute information and / or behavior data from the preprocessed user data to analyze the user's characteristics, such as the user's interest tags, behavior characteristics, time characteristics, etc.

[0047] Before collecting user data, manage user characteristics, including: pre-defining user characteristic attributes, analyzing and storing data according to preset characteristic attributes. For example, for user interest characteristics, establish a user interest table in advance, with fields such as interest name, interest category, preference, and subdivision field, and associate the user ID with the characteristic table for storage, so as to accurately obtain user information through the user ID.

[0048] Step S102, according to the user characteristics of the target user, the preference feature vector of the target user is calculated based on the averaging method and the time decay method.

[0049] Each user has different interests in the same browsing content in different time periods. Combined with the time factor, the preference feature vector of the target user is calculated by the time decay method. The specific steps include:

[0050] 1) Obtain the historical behavior data of the target user and transform the historical behavior data into a multi-dimensional feature vector. The transformed feature vector is used to represent the correspondence between the historical behavior of the target user and the time point.

[0051] The user's historical behavior is recorded in the database. The data in the database records the user's operation records and the time point of the operation. The user's historical behavior data is transformed into a multi-dimensional feature vector matrix. Among them, the dimension of the feature vector matrix is ​​determined based on the number of user operation records. n browsing records correspond to n feature vector matrices.

[0052] Exemplarily, the browsing records of the target user in the past month recorded in the database are obtained. Take 5 browsing records as an example. Each time point corresponds to a browsing record. The 5 browsing records are as follows: browsing record 1 (2023-02-20), browsing record 2 (2023-02-25), browsing record 3 (2023-02-25, 2023-03-02), browsing record 4 (2023-03-02, 2023-03-05), browsing record 5 (2023-03-05, 2023-03-09). These browsing records are converted into a 5-dimensional feature vector matrix, which represents the browsing status of users at different time points for a certain content. The 5-dimensional feature vector can be expressed as: Each row of the matrix corresponds to a browsing record of the user, each column indicates the time corresponding to the browsing record of the user, and each value in the matrix indicates the browsing status of the content corresponding to the browsing record at a specific time point. For example, the value 1 in the first row and the first column of the matrix indicates that the user browsed the content corresponding to the browsing record 1 at the time point of 2023-02-20.

[0053] 2) Based on the multidimensional feature vector, the time decay weight of the feature vector corresponding to the historical behavior data at each time point is calculated.

[0054] Each historical behavior of the user corresponds to a weight, which can be used to quantify the user's interest, preference, and demand for content. For example, the weight can be the user's interest index, preference index, or demand for the content, etc., and can also be understood as the credibility of user characteristics when analyzing the user's preferences or the probability that the user is interested in the content, etc. Specifically, the real-time update of the weight can be to update the weight corresponding to the user's historical behavior in real time by analyzing the user behavior when the user behavior related to the user's user characteristics is detected. Specifically, the weight of the feature vector corresponding to the historical behavior at each time point decays according to an exponential function, and the time decay weight of the feature vector corresponding to each historical behavior is calculated. For example, taking the current time as 2023-02-27 as an example, the time point corresponding to browsing record 1 is 2023-02-20, indicating that the time interval is 7 days, then the calculation of the weight of the feature vector corresponding to this browsing record can be expressed as .

[0055] 3) Based on the average method, the product of the time decay weight and the feature vector at the corresponding time point is calculated, and the product result is used as the preference feature vector of the target user.

[0056] Specifically, based on the average method, the average value of the vectors corresponding to all the user's favorite contents is directly used as the user's preference feature vector. It can be expressed as: Where N represents the total number of content that the user has liked in the past. The feature vector representing the i-th content that the user likes. The final result is the average value of each preference feature. However, in actual scenarios, the user's preference for content decays over time, and the calculation of the user's preference feature vector needs to consider the time factor. Therefore, the product of the time decay weight and the feature vector at the corresponding time point is calculated based on the average method, which can be expressed as: Where N represents the total number of contents in the recorded user's historical behaviors; The feature vector representing the i-th content that the user likes; α is the attenuation factor, which takes a value less than 1. It is the time interval from the time when the user takes action on the i-th content to the current time.

[0057] Step S103: extracting the attribute features of the content to be recommended, and calculating the similarity between the features of the content to be recommended in combination with the preference feature vector of the target user.

[0058] The features of the content include the attribute features of the content itself and the preference features of the user. When calculating the similarity between the content features, the similarity between the attribute features and the similarity between the preference features of the user are calculated respectively.

[0059] Specifically, the step of calculating the similarity between the contents to be recommended includes:

[0060] 1) Extract the attribute features of the content to be recommended and generate the target feature vector;

[0061] 2) Obtain the preference feature vector of the target user;

[0062] 3) Calculate the similarity between target feature vectors and preference feature vectors respectively, and sum them up to get the similarity between the features of the content to be recommended.

[0063] Among them, the similarity is calculated using the cosine similarity algorithm.

[0064] In the above steps S102 and S103, after extracting the user features and the attribute features of the content, the bag-of-words model is used to convert the text features into feature vectors.

[0065] Step S104, based on the preference feature vector of the target user and the similarity between the features of the content to be recommended, the matching degree between the target user and each content to be recommended is calculated to generate a recommendation list.

[0066] Specifically, the matching degree between the target user and each content to be recommended is calculated, which can be expressed as: represents the target user's preference feature vector, is the attribute feature vector of the content to be recommended, k is the kth feature, and there are K features in total. The closer the calculated value is to 1, the more similar the two vectors are. Specifically, after calculating the preference feature vector of the target user, the content object corresponding to the preference feature vector can be determined from the historical behavior data of the target user, and then according to the similarity between the contents, the content to be recommended for the user can be determined, and the matching degree between the target user and each content to be recommended is calculated, and each matching degree is sorted to generate a recommendation list. According to the sorting result in the recommendation list, the target content corresponding to the target matching degree is determined, and the target content is recommended to the target customer. Among them, if the sorting result is a descending sorting result, a preset number of contents ranked first can be selected as the target content.

[0067] In implementation, taking browsing records as an example, based on the user's preference feature vector, the browsing content in the user's historical behavior data can be determined first, and then according to the similarity between the content features calculated in step S103, the browsing content in the user's historical behavior can be screened out in the database for content similar to the user's historical browsing content, and the screened similar content can be used as content to be recommended. Finally, the similarity between the user's preference features and the attribute features of each content to be recommended is calculated, and a recommendation list is generated according to the similarity.

[0068] Figure 2This is a schematic diagram of the overall framework of an intelligent recommendation method for government services provided in an embodiment of the present invention. The intelligent recommendation method for government services of the present invention adopts the Apache Flink framework, which mainly includes a data layer, an application layer, an interface layer, and a display layer. Among them, the data layer connects to external data sources, collects users' basic data, behavior data, and portrait data, and performs structured storage. The application layer is used to perform operations such as booth configuration, user feature management, crowd management, service management, algorithm management, and polling services to calculate user recommended content. The interface layer connects to the government applets. The display layer is used to display data and content on each page of the government applets.

[0069] When recommending content to users, we monitor users’ behavior data and contextual information in real time, use real-time data processing technology to analyze and process user behavior data in real time, and dynamically adjust recommendation strategies and personalized push content based on users’ real-time needs and personalized characteristics to provide real-time and adaptive push strategies.

[0070] Specifically, the use of real-time data processing technology to process user behavior data mainly includes: data stream input, data preprocessing and feature extraction. Among them, through load balancing technology, network traffic is shared to multiple servers or links, improving network throughput and response speed, thereby reducing network congestion. At the same time, the transmitted data is compressed to reduce the amount of data transmission and speed up data transmission, thereby alleviating network congestion problems.

[0071] The embodiment of the present invention determines the recommended content that the target user may be interested in based on the user characteristics of the target user, calculates the matching degree between the target user and each recommended content, and then determines the target content that can be recommended to the target user. It can provide the user with personalized recommendation services based on the user's needs, preferences and historical behaviors, improves the accuracy of the matching degree calculation between the user and the recommended content, and then improves the accuracy of recommending content to the user.

[0072] Figure 3 Schematic diagram of a structure of a government service intelligent recommendation system provided by an embodiment of the present invention. The system is generally configured in a recommendation device and can intelligently recommend services to users by executing a government service intelligent recommendation method. Figure 3 As shown, the system includes:

[0073] A user feature acquisition module is used to collect multi-dimensional data of target users and pre-process the data to obtain user features of target users;

[0074] A preference feature vector generation module is used to calculate the preference feature vector of the target user based on the user characteristics of the target user and the averaging method and the time decay method;

[0075] A similarity calculation module is used to extract the attribute features of the content to be recommended and calculate the similarity between the features of the content to be recommended in combination with the preference feature vector of the target user;

[0076] The recommendation list generation module is used to calculate the matching degree between the target user and each content to be recommended based on the preference feature vector of the target user and the similarity between the features of the content to be recommended, and generate a recommendation list.

[0077] Figure 4 is a schematic diagram of a recommended device provided by an embodiment of the present invention, such as Figure 4 As shown, the recommendation device includes a processor 400, a memory 410, an input device 420, and an output device 430; the number of the processor 400 in the recommendation device can be one or more. Figure 4 A processor 400 is taken as an example; the processor 400, the memory 410, the input device 420 and the output device 430 in the recommended device can be connected via a bus or other means. Figure 4 The example of connecting through bus is taken in the following.

[0078] The memory 410, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as program instructions and / or modules corresponding to the recommendation method in the embodiment of the present invention (for example, the user feature acquisition module 310, the preference feature vector generation module 320, the similarity calculation module 330 and the recommendation list generation module 340). The processor 400 executes the various functional applications and data processing of the recommendation device by running the software programs, instructions and modules stored in the memory 410, that is, implements the above-mentioned recommendation method.

[0079] The memory 410 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 410 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 410 may further include a memory remotely arranged relative to the processor 400, and these remote memories may be connected to the recommended device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0080] The input device 420 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the recommended device. The output device 430 may include a display device such as a display screen.

[0081] An embodiment of the present invention further provides a storage medium containing computer executable instructions, and when the computer executable instructions are executed by a computer processor, they are used to perform a recommendation method, the method comprising:

[0082] Collect multi-dimensional data of target users and pre-process the data to obtain user characteristics of target users;

[0083] According to the user characteristics of the target user, the preference feature vector of the target user is calculated based on the average method and the time decay method;

[0084] Extract the attribute features of the content to be recommended, and calculate the similarity between the features of the content to be recommended based on the preference feature vector of the target user;

[0085] Based on the preference feature vector of the target user and the similarity between the features of the content to be recommended, the matching degree between the target user and each content to be recommended is calculated to generate a recommendation list.

[0086] Of course, a storage medium containing computer executable instructions provided in an embodiment of the present invention, whose computer executable instructions are not limited to the above-mentioned method operations, can also perform the relevant operations in the recommended method provided in any embodiment of the present invention. Through the above description of the implementation mode, the technicians in the relevant field can clearly understand that the present invention can be implemented with the help of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation mode. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer's floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform the method of the embodiment of the present invention.

[0087] It is worth noting that in the embodiments of the above-mentioned recommended device, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0088] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for intelligent recommendation of government services, characterized in that: include: Collecting multi-dimensional data of target users, and preprocessing the data to obtain user characteristics of the target users; According to the user characteristics of the target user, a preference feature vector of the target user is calculated based on an averaging method and a time decay method; Determine the content object corresponding to the preference feature vector from the historical behavior data of the target user, and then determine the content to be recommended to the user based on the similarity between the contents; The method of calculating the target user's preference feature vector based on the user characteristics of the target user and based on the averaging method and the time decay method includes: obtaining the target user's historical behavior data and transforming the historical behavior data into a multidimensional feature vector; calculating the time decay weight of the feature vector corresponding to the historical behavior data at each time point based on the multidimensional feature vector; calculating the product result of the time decay weight and the feature vector at the corresponding time point based on the averaging method, and taking the product result as the target user's preference feature vector; the weight of the feature vector corresponding to the historical behavior at each time point decays according to an exponential function, and calculating the time decay weight of the feature vector corresponding to each historical behavior; The step of obtaining the historical behavior data of the target user and transforming the historical behavior data into a multidimensional feature vector includes: obtaining the operation record of the target user recorded in the historical behavior data and the time point of performing the operation; constructing the historical behavior data of the target user into a multidimensional feature vector matrix based on the operation record and the time point of performing the operation; and determining the dimension of the feature vector matrix based on the number of operation records; The product result of the time decay weight and the feature vector at the corresponding time point calculated based on the averaging method can be expressed as: N represents the total number of contents in the recorded user's historical behavior; w i The feature vector representing the i-th content that the user likes; α is the attenuation factor, which takes a value less than 1, diff i The time interval from when the user takes action on the i-th content to the current time; Content features include the attribute features of the content itself and the preference features of the user. When calculating the similarity between content features, the similarity between the attribute features and the similarity between the preference features of the user are calculated separately, and the similarity between the content features is obtained by summing the two. Based on the user's preference feature vector, the browsing content in the user's historical behavior data is determined, and according to the calculated similarity between the content features, the content similar to the user's historical browsing content is screened out in the database through the browsing content in the user's historical behavior, and the screened similar content is used as the content to be recommended; the matching degree between the target user and each content to be recommended is calculated, and a recommendation list is generated; Before collecting the multi-dimensional data of the target user, the user characteristics of the target user are managed, including: pre-defining the characteristic attributes of the user, analyzing and storing the data according to the preset characteristic attributes; for the user's interest characteristics, establishing a user interest table in advance, and the table fields include interest name, interest classification, preference level, and subdivision field.

2. According to claim 1, a method for intelligent recommendation of government services is characterized in that: The calculation of the matching degree between the target user and each content to be recommended can be expressed as: in, represents the target user's preference feature vector, is the attribute feature vector of the content to be recommended, k represents the kth feature, and there are K features in total.

3. An intelligent government service recommendation system, characterized in that: include: A user feature acquisition module, used to collect multi-dimensional data of target users and pre-process the data to obtain user features of the target users; A preference feature vector generation module, used to calculate the preference feature vector of the target user based on the user characteristics of the target user and the averaging method and the time decay method; The method of calculating the target user's preference feature vector based on the user characteristics of the target user and based on the averaging method and the time decay method includes: obtaining the target user's historical behavior data and transforming the historical behavior data into a multidimensional feature vector; calculating the time decay weight of the feature vector corresponding to the historical behavior data at each time point based on the multidimensional feature vector; calculating the product result of the time decay weight and the feature vector at the corresponding time point based on the averaging method, and taking the product result as the target user's preference feature vector; the weight of the feature vector corresponding to the historical behavior at each time point decays according to an exponential function, and calculating the time decay weight of the feature vector corresponding to each historical behavior; The step of obtaining the historical behavior data of the target user and transforming the historical behavior data into a multidimensional feature vector includes: obtaining the operation record of the target user recorded in the historical behavior data and the time point of performing the operation; constructing the historical behavior data of the target user into a multidimensional feature vector matrix based on the operation record and the time point of performing the operation; and determining the dimension of the feature vector matrix based on the number of operation records; The product result of the time decay weight and the feature vector at the corresponding time point calculated based on the averaging method can be expressed as: N represents the total number of contents in the recorded user's historical behavior; w i The feature vector representing the i-th content that the user likes; α is the attenuation factor, which takes a value less than 1, diff i The time interval from when the user takes action on the i-th content to the current time; A similarity calculation module is used for content features including attribute features of the content itself and preference features of the user. When calculating the similarity between content features, the similarity between attribute features and the similarity between user preference features are calculated respectively, and the similarity between the attribute features and the similarity between the user preference features are summed to obtain the similarity between the content features; The recommendation list generation module is used to determine the browsing content in the user's historical behavior data based on the user's preference feature vector, and filter out the content similar to the user's historical browsing content in the database through the browsing content in the user's historical behavior according to the calculated similarity between the content features, and use the filtered similar content as the content to be recommended; calculate the matching degree between the target user and each content to be recommended, and generate a recommendation list; Before collecting the multi-dimensional data of the target user, the user characteristics of the target user are managed, including: pre-defining the characteristic attributes of the user, analyzing and storing the data according to the preset characteristic attributes; for the user's interest characteristics, establishing a user interest table in advance, and the table fields include interest name, interest classification, preference level, and subdivision field.

4. A recommended device, characterized in that: include: one or more processors; A memory for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the government service intelligent recommendation method as described in any one of claims 1 to 2.

5. A storage medium containing computer executable instructions, characterized in that: The computer executable instructions are used to execute the government service intelligent recommendation method as described in any one of claims 1 to 2 when executed by a computer processor.