An optimal search method for integrated digital resource systems

Abstract

The present invention provides an optimal search method applied to an integrated digital resource system, and belongs to the field of intelligent search technology. The present invention provides an optimal search method for digital resources when the number of users is large, performs label and portrait analysis on system users, and uses a naive Bayes algorithm to analyze and search for users' interest data and behavior data to obtain the optimal resources, thereby effectively improving the use efficiency and experience of various levels and departments.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent search of big data, and in particular to an optimal search method applied to an integrated digital resource system. Background Art

[0002] In the process of promoting smart city construction, in the past, in the process of smart city informatization construction in various places, there were thousands of digital resources such as various data, systems, components, etc. scattered in various places and departments. In order to solve this problem, an integrated intelligent public data platform came into being.

[0003] The IRS system integrates a large amount of digital resources such as digital applications, public data, and intelligent components, classifies and catalogs various data resources, and creates an account of urban resources. Using technologies such as intelligent retrieval and graph knowledge reasoning, with the help of the constructed digital resource general ledger, users can browse data resources in one stop, making resource management more intelligent and automated. Through the construction of the IRS system, the problem of low-level duplication of urban informatization can be effectively solved, and the data sharing guarantee capability and efficiency can be improved.

[0004] After the system is built, digital resources will become more dense, and unfiltered information will have poor readability, making the system less user-friendly. Therefore, it is particularly important to refine and filter information so that the presented information matches the user's needs, knowledge structure, and purpose. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides an optimal search method for an integrated digital resource system. Using big data tags and portrait analysis technology, personnel and resources are modeled and calculated to analyze the resources that are useful or interesting to users. The naive Bayes algorithm is used to analyze the user's interest data and behavior data to search for the optimal resources, greatly improving the user's system experience and efficiency.

[0006] The technical solution of the present invention is:

[0007] An optimal search method applied to an integrated digital resource system. During the IRS system construction phase, a user portrait is formulated based on user type, department, responsibility, etc., combined with user basic information, self-attached interest tags, and behavior analysis of users at the same level. When the user enters a keyword to search for resources, the system performs natural semantic analysis on the keyword and selects the corresponding association analysis method. When there are a large number of users, the naive Bayes algorithm is used to analyze and predict the ranking of resources that the user is interested in, and the user performs subsequent usage operations based on the feedback results.

[0008] Furthermore,

[0009] The specific steps are as follows:

[0010] Step 1) Plan user portraits and set user tags;

[0011] Step 2) When the user submits a search operation, it is determined whether there is behavior data of other users in the same layer, and the correlation analysis method is selected;

[0012] Step 3), perform natural semantic analysis of the search keywords and perform association analysis according to the definition in step 2);

[0013] Step 4) Based on the calculated association ranking results, the results are displayed and output on the system interface, and the user performs corresponding subsequent operations.

[0014] Going further,

[0015] In step 1), plan the user basic information table (user_infor), interest tag table (prefer_infor), and user behavior table (user_behavior_infor).

[0016] The user basic information table maintains basic information such as user name, type, and contact information. The interest tag table maintains user department, responsibilities, and positions. The user behavior table associates user information and interest tags for subsequent association analysis of user search recommendations.

[0017] In the interest tag table, the correspondence between user responsibilities and positions is established, and users select their departments and positions when they register for the first time;

[0018] According to the interest tag table, match the department number corresponding to the user as the basis for user stratification;

[0019] Establish the relationship between responsibility labels and resource categories;

[0020] Based on the user stratification results, role matching is performed, and users select their own "interest" tags from the matching roles.

[0021] Going further,

[0022] In step 2),

[0023] (1) If there is no behavioral data of other users in the same layer, the association ranking result is obtained based on the method of "basic information of the user + self-attached interest tags" during the association analysis;

[0024] (2) If there is behavioral data of other users in the same layer, the association ranking result is obtained based on the method of "basic information of the user + self-attached interest tags, and behavioral data of other users in the same layer" during the association analysis.

[0025] Going further,

[0026] In step 3),

[0027] (1) According to the method of "basic information of the user + self-attached interest tags" in step 2), if there is no behavioral data of users in the same layer, or the number of users in the same layer is small, the resource categories corresponding to the responsibility tag will cast 1 vote each, and so on. In the end, all resource categories have their own "votes", that is, the ranking order.

[0028] (2) Based on the method of "basic information of the user + self-attached interest tags, and behavioral data of other users in the same layer" in step 2), under the premise of a large number of users, the naive Bayes algorithm is used to calculate the decomposition.

[0029] The decomposition is calculated using the Naive Bayes algorithm, including:

[0030] 1) Solve according to the naive Bayes formula and perform correlation analysis calculation;

[0031] 2) Plan and design the user behavior table and update the association analysis results based on the calculation results;

[0032] 3) For user N, the ranking result of the concerned categories is calculated;

[0033] Finally, the analysis result of the user is obtained, and the analysis shows which category of products the user focuses on.

[0034] 4) On this basis, a secondary analysis is conducted to obtain the usage ratio of different types of resources according to the usage history of users at the same level as user N. The relevance ranking of user N is obtained by matrix decomposition, and the ranking of the user in each category is obtained.

[0035] The beneficial effects of the present invention are

[0036] By using big data tags and portrait analysis technology during the system design phase, modeling and calculating personnel and resources, analyzing the resources that are useful or interesting to users, and using the naive Bayes algorithm to analyze the user's interest data and behavior data to search for the optimal resources, the user's system experience and efficiency are greatly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic diagram of the workflow of the present invention. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0039] The present invention provides an optimal search method applied to an integrated digital resource system, which uses big data tags and portrait analysis technology to model and calculate personnel and resources, analyzes resources that are useful or interesting to users, and uses a naive Bayes algorithm to analyze users' interest data and behavior data to search for the optimal resources.

[0040] During the IRS system construction phase, user portraits are developed based on user types, departments, responsibilities, etc., combined with user basic information, self-attached interest tags, and behavior analysis of users at the same level. When users enter keywords to search for resources, the system performs natural semantic analysis on the keywords and selects the corresponding association analysis method for calculation. For a large number of users, the naive Bayes algorithm is used to analyze and predict the ranking of resources that users are interested in, and users perform subsequent usage operations based on the feedback results.

[0041] To achieve the above purpose, the specific implementation methods are as follows:

[0042] Step 1: Plan user profiles and set user tags. Plan user basic information table

[0043] (user_infor), interest tag table (prefer_infor), user behavior table (user_behavior_infor). The user basic information table is used to maintain basic information such as user name, type, contact information, etc. The interest tag table is used to maintain information such as user department, responsibilities, and position. The user behavior table associates user information and interest tags for subsequent association analysis of user search recommendations.

[0044] (1) In the interest tag table, establish the corresponding relationship between user responsibilities, departments, etc., and users select departments and positions when they first register. The following table is an example table of city interest setting relationships, where the interest setting tags include: city overview, economic development, ecological livability, construction leadership, cultural tourism, government services, safe production, transportation operation, educational resources, medical health, people's livelihood security, city management, cloud network number, city infrastructure, etc. The example table is replaced by responsibility 1, responsibility 2, responsibility 3, etc.

[0045] department Department Number Responsibilities A Committee 1 Responsibility 1, Responsibility 4, Responsibility 5, Responsibility 6, Responsibility 8, Responsibility 10, Responsibility 11, Responsibility 14 Bureau B 2 Responsibility 2, Responsibility 3, Responsibility 5, Responsibility 7, Responsibility 8, Responsibility 10 Bureau C 3 Responsibility 1, Responsibility 2, Responsibility 4, Responsibility 6, Responsibility 7, Responsibility 9, Responsibility 10, Responsibility 13 D 4 Responsibility 1, Responsibility 2, Responsibility 3, Responsibility 4, Responsibility 7, Responsibility 10, Responsibility 12, Responsibility 14 F Bureau 5 Responsibility 1, Responsibility 4, Responsibility 5, Responsibility 6, Responsibility 8, Responsibility 10, Responsibility 11, Responsibility 15 G Bureau 6 Responsibility 2, Responsibility 4, Responsibility 5, Responsibility 6, Responsibility 8, Responsibility 10, Responsibility 11, Responsibility 14, Responsibility 14 …… …… …… W Bureau N Responsibility 3, Responsibility 4, Responsibility 5, Responsibility 7, Responsibility 8, Responsibility 10, Responsibility 11, Responsibility 12

[0046] Example table of interest setting relationships

[0047] (2) According to the interest-based relationship example table, the department number corresponding to the user is matched as the basis for user stratification.

[0048] (3) Establish the relationship between responsibility labels and resource categories. There is a mapping relationship between interest labels and resource categories. Each responsibility may correspond to multiple related resource categories, including application resources, data resources, component resources, cloud network resources, and perception resources.

[0049] For example, "Responsibility 1" corresponds to "application resources, data resources", and "Responsibility 2" corresponds to "data resources, cloud network resources".

[0050] (4) Based on the user stratification results, the user matches the responsibilities and selects his or her own "interest" tags from the matching responsibilities. For example, if the user belongs to department A, the selected interest tags are: responsibility 1, responsibility 5, responsibility 8, responsibility 11, and responsibility 15.

[0051] Step 2: When the user submits a search operation, determine whether there is behavior data of other users at the same level and select a correlation analysis method.

[0052] (1) If there is no behavioral data from other users in the same layer, the association ranking results are obtained based on the "basic information of the user + self-attached interest tags" method during the association analysis.

[0053] (2) If there is behavioral data of other users in the same layer, the association ranking result is obtained based on the method of "basic information of the user + self-attached interest tags, and behavioral data of other users in the same layer" during the association analysis.

[0054] Step 3: Perform natural semantic analysis of the search keywords and conduct association analysis based on the definition in step 2.

[0055] (1) According to the method of "basic information of the user + self-attached interest tags" in step 2, if there is no behavioral data of users in the same layer, or the number of users in the same layer is small, preferably, the resource categories corresponding to the responsibility tag will cast 1 vote each, and so on. Finally, all resource categories have their own "number of votes", that is, the ranking order.

[0056] (2) Based on the method in step 2 based on "the user's basic information + self-attached interest tags, and the behavioral data of other users in the same layer", under the premise that there are a large number of users, the naive Bayes algorithm is used to calculate the decomposition.

[0057] The Naive Bayes algorithm is: P(C|F1 F2…Fn)=P(F1 F2…Fn|C)*P(C) / P(F1 F2…Fn), where F is the feature and C is the classification. F is (department, position, level, self-labeled “interest” label responsibility).

[0058] 1) Solve according to the naive Bayes formula and perform association analysis calculations.

[0059]

[0060] Where F is the feature and C is the classification. We get:

[0061]

[0062] Similarly, we can ask:

[0063] P2 = P (responsibility 2 | department, position, level, label responsibility)

[0064] P3 = P (responsibility 3 | department, position, level, label responsibility)

[0065] P4 = P (responsibility 4 | department, position, level, label responsibility)

[0066] 2) Plan and design the user behavior table, and update the association analysis results based on the calculation results.

[0067]

[0068] User behavior concept table

[0069] 3) For user N, the ranking results of the concerned categories are calculated.

[0070]

[0071]

[0072] The Class() function is used to obtain the corresponding category, and the CStr() function is used to convert it into a string.

[0073] Finally, the analysis results of the user are obtained, and the type of products that the user focuses on are analyzed.

[0074] index_all is the demand coefficient arranged from large to small [responsibility 1, responsibility 2, responsibility 3, responsibility 4].

[0075] 4) On this basis, a secondary analysis is conducted to obtain the usage ratio of different types of resources according to the usage history of users at the same level as user N. The relevance ranking of user N is obtained by matrix decomposition, and the ranking of the user in each category is obtained.

[0076] Step 4: Based on the calculated association ranking results, the results are displayed and output on the system interface, and the user performs corresponding follow-up operations.

[0077] The IRS system is positioned to build an intelligent account book for government data resources, integrating relatively discrete application systems, capability components, data, cloud networks, perception and other data resources in the city into a comprehensive organic whole for users to search, browse and apply for use. The present invention can achieve the optimal search for digital resources when the number of users is large, aiming to improve the user's efficiency and experience in the system, and effectively avoid the problem of resources not matching user needs.

[0078] The above description is only a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, and is not used to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. An optimal search method applied to an integrated digital resource system, characterized in that: Combining basic user information, self-posted interest tags, and behavior analysis of users at the same level, a user portrait is developed. When a user enters a keyword to search for resources, the system performs natural semantic analysis on the keyword and selects the corresponding association analysis method. When the number of users at the same level is not less than 50, the naive Bayes algorithm is used to analyze and predict the ranking of resources that the user is interested in. The user performs subsequent operations based on the feedback results. The specific steps are as follows: Step 1) Plan user portraits and set user tags; Step 2), when the user submits a search operation, determine whether there is behavioral data of other users on the same level and select the association analysis method; Step 3), perform natural semantic analysis of the search keywords and perform association analysis according to the definition in step 2); Step 4), based on the calculated correlation ranking results, the results are displayed and output on the system interface, and the user performs corresponding follow-up operations; In step 1), plan the user basic information table (user_infor), interest tag table (prefer_infor), and user behavior table (user_behavior_infor); In step 2), (1) If there is no behavioral data of other users in the same layer, the association ranking result is obtained based on the "basic information of the user + self-attached interest tags" method during the association analysis; (2) If there is behavioral data of other users in the same layer, the association ranking result is obtained based on the method of "basic information of the user + self-attached interest tags, and behavioral data of other users in the same layer" during the association analysis; In step 3), (1) According to the method of "basic information of the user + self-attached interest tags" in step 2), if there is no behavioral data of users in the same layer, or the number of users in the same layer is less than 50, the resource categories corresponding to the responsibility tags in the interest tags will cast 1 vote each, and so on. In the end, all resource categories will have their own "number of votes", that is, the ranking order; (2) Based on the method of "basic information of the user + self-attached interest tags, and behavioral data of other users in the same layer" in step 2), the naive Bayes algorithm is used to calculate the decomposition, provided that the number of users in the same layer is not less than 50; The decomposition is calculated using the Naive Bayes algorithm, including: 1) Solve according to the naive Bayes formula and perform correlation analysis calculation; 2) Plan and design the user behavior table and update the association analysis results based on the calculation results; 3) For user N, calculate the ranking result of the concerned categories; Finally, the analysis result of the user is obtained, and the analysis shows which category of products the user focuses on. 4) On this basis, a secondary analysis is conducted to obtain the usage ratio of different types of resources according to the usage history of users at the same level as user N. The relevance ranking of user N is obtained by matrix decomposition, and the ranking of the user in each category is obtained.

2. The method according to claim 1, characterized in that The user basic information table maintains the basic information of the user's name, type, and contact information; the interest tag table maintains the user's department, responsibilities, and position; and the user behavior table associates user information and interest tags for subsequent association analysis of user search recommendations.

3. The method according to claim 2, characterized in that In the interest tag table, the correspondence between user responsibilities and positions is established, and users select their departments and positions when they register for the first time; According to the interest tag table, match the department number corresponding to the user as the basis for user stratification; Establish the relationship between responsibility labels and resource categories; Based on the user stratification results, role matching is performed, and users select their own "interest" tags from the matching roles.

Images