Operating service item association retrieval method and system based on portrait model

Abstract

The invention relates to the technical field of searching of government affair service items, in particular to an affair handling service item association retrieval method and system based on a portrait model, and the method comprises an item search portrait model training process and a search process; the method has the beneficial effects that the item search portrait model is established by utilizing information such as search content of a user, a clicked item page and stay duration in a corresponding page, and when the user searches subsequently, the association similarity between the search content and the item content is analyzed through the item search portrait model, so that the item search portrait model is more accurate. The item most relevant to the search content of the user is displayed preferentially, so that the user can find the relevant content more efficiently and conveniently, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of searching for government service items, and specifically to a method and system for related retrieval of government service items based on a portrait model. Background Art

[0002] Currently, most government service websites offer search functionality, but the results are often suboptimal. For example, the desired item may not be found, or the item may be ranked low, requiring scrolling down or turning pages to find it. This results in a poor user experience on government service websites. Summary of the Invention

[0003] The purpose of the present invention is to provide a service item association retrieval method and system based on a portrait model to solve the problems raised in the above background technology.

[0004] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a service item association retrieval method based on a portrait model, comprising a task search portrait model training process and a search process;

[0005] The training process of the item search portrait model includes: performing item search; segmenting the search keywords; recording the items clicked by users and the length of time they stay on the corresponding pages; calculating the similarity between the clicked items and the search keywords; and training the item search portrait model based on the above data.

[0006] The search process includes: conducting an item search; segmenting the search keywords; directly using the keywords to perform fuzzy matching on the items, and recording the results as result 1; calculating the correlation similarity between the keywords and all items through the item search portrait model, and outputting result 2 after sorting by correlation similarity; integrating result 1 and result 2 and returning the final result.

[0007] Preferably, the word segmentation library for the search content adopts IKAnalyzer.

[0008] Preferably, the matter search portrait model is implemented using the KNN algorithm, and matters are classified by measuring the distance between different eigenvalues.

[0009] Preferably, in the training process of the item search portrait model, the keywords searched by the user, the items clicked subsequently, and the length of time the user stays on the page after opening it are recorded. Based on this information, the correlation similarity between the clicked items and the search keywords is calculated to establish the item search portrait model.

[0010] Preferably, in the search process, the result 1 obtained by directly using keywords for fuzzy matching of matters and the result 2 obtained after calculating the correlation similarity ranking through the matter search portrait model are integrated and processed. The integration method is to comprehensively consider the two results, so that the matters most relevant to the user's search content are displayed first, and finally the integrated results are returned.

[0011] A system for service item association retrieval method based on a portrait model, comprising an item search portrait model training module and a search processing module;

[0012] The item search portrait model training module is used to execute the item search portrait model training process, which includes: performing item search operations; performing word segmentation processing on search keywords; recording items clicked by users and the length of time they stay on the corresponding pages; calculating the similarity between the clicked items and the search keywords; and training the item search portrait model based on the above data;

[0013] The search processing module is used to execute a search process, which includes: conducting an item search; segmenting the search keywords; directly using the keywords to perform fuzzy matching on the items, and recording the matching results as result 1; calculating the correlation similarity between the keywords and all items through the item search portrait model, and outputting result 2 after sorting by correlation similarity; integrating result 1 and result 2, and returning the integrated result as the final result.

[0014] Preferably, in the search processing module, the word segmentation library for the search content adopts IKAnalyzer.

[0015] Preferably, in the matter search portrait model training module, the matter search portrait model is implemented using the KNN algorithm, and matters are classified by measuring the distance between different eigenvalues.

[0016] Preferably, the item search portrait model training module records the search keywords, the items clicked subsequently, and the length of time the user stays on the page when searching, and calculates the correlation similarity between the clicked items and the search keywords based on this information to establish an item search portrait model.

[0017] Preferably, in the search processing module, the result 1 obtained by directly using keywords for fuzzy matching of matters and the result 2 obtained after calculating the correlation similarity and sorting through the matter search portrait model are integrated and processed. The integration processing method is to comprehensively consider the two results, so that the matters most relevant to the user's search content are displayed first, and finally the integrated results are returned to improve the efficiency and convenience of users in finding related content.

[0018] Compared with the prior art, the present invention has the following beneficial effects:

[0019] The present invention proposes a service item association retrieval method and system based on a portrait model, which uses information such as previous user search content, clicked item pages, and length of stay on the corresponding page to establish an item search portrait model. When subsequent users search, the item search portrait model is used to analyze the correlation similarity between the search content and the item content, so that the items most relevant to the user's search content are displayed first, allowing users to find relevant content more efficiently and conveniently, thereby improving user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A flowchart for searching for a portrait model training process for the present invention;

[0021] Figure 2 Search flowchart for the present invention. DETAILED DESCRIPTION

[0022] In order to clearly and completely describe the objectives and technical solutions of the present invention and make the advantages more clearly understood, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are part of the embodiments of the present invention, not all of them, and are only used to explain the embodiments of the present invention, not to limit the embodiments of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0023] For example 1, please refer to Figures 1 to 2 The present invention provides a technical solution: a service item related retrieval method based on a portrait model. First, a service item search portrait model is established using information such as previous user search content, clicked item pages, and length of stay on the corresponding page. When subsequent users search, the item search portrait model is used to analyze the correlation similarity between the search content and the item content, so that the items most relevant to the user's search content are displayed first, allowing users to find relevant content more efficiently and conveniently, thereby improving user experience.

[0024] Includes the event search portrait model training process and search process.

[0025] The training process of the event search portrait model includes:

[0026] 1) Search for items,

[0027] 2)Search keyword segmentation,

[0028] 3) Record user clicks and page stay time

[0029] 4) Calculate the similarity between click items and search keywords

[0030] 5) Training the event search portrait model

[0031] The search process includes:

[0032] 1) Item Search

[0033] 2) Search keyword segmentation

[0034] 3) Directly use keywords for fuzzy matching

[0035] 4) Calculate the correlation similarity between keywords and all matters through the matter search portrait model

[0036] 5) Return the processing results after comprehensive processing

[0037] Furthermore, the search content segmentation database uses IKAnalyzer, and the item search profiling model is implemented using the KNN algorithm. This classifies items by measuring the distance between different feature values. When a user searches, the search keyword, the item clicked, and the duration of the page after opening are recorded. The correlation similarity between the clicked item and the search keyword is calculated to establish the item search profiling model.

[0038] Furthermore, after the model is established, when the user searches, the search keywords are first used to directly match the items, and the result is recorded as result 1. Secondly, all keywords are input into the item search portrait model, and result 2 is output after sorting by the correlation similarity between the keywords and the items. The data of result 1 and result 2 are integrated and processed to return the final result.

[0039] Figure 1 It is a flowchart for implementing the training of the item search portrait model. Its implementation content includes: first, collecting model training information, mainly including the user's historical search information and the clicks and stay time on the corresponding page after the search. For example, after the user searches for the keyword "personal household registration", it is first segmented into "individual" and "household registration" through the IKAnalyzer word segmentation library, and at the same time, it is recorded in the returned result list that "household registration matter certificate" and "sole proprietorship establishment matter" have been clicked, and the stay time on the opened page is 5 seconds and 1 minute respectively. Then, the KNN algorithm is used to build a model, and classification is performed by measuring the distance between different feature values to establish an item search portrait model.

[0040] Figure 2 This is a flowchart for item search. Its implementation content includes: after the user enters the search content "personal household registration", it first directly matches the item content and records it as result one. Then, the IKAnalyzer word segmentation library segments "personal" and "household registration". The item search portrait model is used to calculate the correlation similarity between "personal" and "household registration" and the search content and different items. The items are sorted according to the similarity, and the sorted results are recorded as result two. The contents of result one and result two are merged and returned.

[0041] In the second embodiment, based on the first embodiment, a system for service item association retrieval method based on a portrait model is proposed, which includes an item search portrait model training module and a search processing module;

[0042] The item search portrait model training module is used to execute the item search portrait model training process, which includes: performing item search operations; performing word segmentation processing on search keywords; recording items clicked by users and the length of time they stay on the corresponding pages; calculating the similarity between the clicked items and the search keywords; and training the item search portrait model based on the above data;

[0043] The search processing module is used to execute a search process, which includes: conducting an item search; segmenting the search keywords; directly using the keywords to perform fuzzy matching on the items, and recording the matching results as result 1; calculating the correlation similarity between the keywords and all items through the item search portrait model, and outputting result 2 after sorting by correlation similarity; integrating result 1 and result 2, and returning the integrated result as the final result.

[0044] In the search processing module, the word segmentation library of the search content uses IKAnalyzer.

[0045] In the matter search portrait model training module, the matter search portrait model is implemented using the KNN algorithm, and matters are classified by measuring the distance between different feature values.

[0046] When a user searches, the item search portrait model training module records the search keywords, the items clicked subsequently, and the length of time the user stays on the page after opening it. Based on this information, the correlation similarity between the clicked items and the search keywords is calculated to establish an item search portrait model.

[0047] In the search processing module, the result 1 obtained by directly using keywords for fuzzy matching of matters and the result 2 obtained by calculating the correlation similarity and sorting through the matter search portrait model are integrated and processed. The integration processing method is to comprehensively consider the two results, so that the matters most relevant to the user's search content are displayed first, and finally the integrated results are returned to improve the efficiency and convenience of users in finding related content.

[0048] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A service item association retrieval method based on a portrait model, characterized by: Including the event search portrait model training process and search process; The training process of the item search portrait model includes: performing item search; segmenting the search keywords; recording the items clicked by users and the length of time they stay on the corresponding pages; calculating the similarity between the clicked items and the search keywords; and training the item search portrait model based on the above data. The search process includes: conducting an item search; segmenting the search keywords; directly using the keywords to perform fuzzy matching on the items, and recording the results as result 1; calculating the correlation similarity between the keywords and all items through the item search portrait model, and outputting result 2 after sorting by correlation similarity; integrating result 1 and result 2 and returning the final result.

2. The service item related retrieval method based on the portrait model according to claim 1 is characterized by: The word segmentation library for search content uses IKAnalyzer.

3. The service item related retrieval method based on the portrait model according to claim 2 is characterized by: The event search portrait model is implemented using the KNN algorithm, which classifies events by measuring the distance between different feature values.

4. The service item related retrieval method based on the portrait model according to claim 3 is characterized by: In the training process of the item search portrait model, the keywords searched by the user, the items clicked subsequently, and the length of time the user stays on the page after opening it are recorded. Based on this information, the correlation similarity between the clicked items and the search keywords is calculated to establish the item search portrait model.

5. The service item related retrieval method based on the portrait model according to claim 4 is characterized by: In the search process, the result 1 obtained by directly using keywords for fuzzy matching of matters and the result 2 obtained by calculating the correlation similarity and sorting through the matter search portrait model are integrated and processed. The integration method is to comprehensively consider the two results, so that the matters most relevant to the user's search content are displayed first, and finally the integrated results are returned.

6. A system for use in the service item association retrieval method based on the portrait model according to claim 5, characterized in that: Includes event search portrait model training module and search processing module; The item search portrait model training module is used to execute the item search portrait model training process, which includes: performing item search operations; performing word segmentation processing on search keywords; recording items clicked by users and the length of time they stay on the corresponding pages; calculating the similarity between the clicked items and the search keywords; and training the item search portrait model based on the above data; The search processing module is used to execute a search process, which includes: conducting an item search; segmenting the search keywords; directly using the keywords to perform fuzzy matching on the items, and recording the matching results as result 1; calculating the correlation similarity between the keywords and all items through the item search portrait model, and outputting result 2 after sorting by correlation similarity; integrating result 1 and result 2, and returning the integrated result as the final result.

7. A system according to claim 6, characterized in that: In the search processing module, the word segmentation library of the search content uses IKAnalyzer.

8. A system according to claim 7, characterized in that: In the matter search portrait model training module, the matter search portrait model is implemented using the KNN algorithm, and matters are classified by measuring the distance between different feature values.

9. A system according to claim 8, characterized in that: When a user searches, the item search portrait model training module records the search keywords, the items clicked subsequently, and the length of time the user stays on the page after opening it. Based on this information, the correlation similarity between the clicked items and the search keywords is calculated to establish an item search portrait model.

10. A system according to claim 9, characterized in that: In the search processing module, the result 1 obtained by directly using keywords for fuzzy matching of matters and the result 2 obtained by calculating the correlation similarity and sorting through the matter search portrait model are integrated and processed. The integration processing method is to comprehensively consider the two results, so that the matters most relevant to the user's search content are displayed first, and finally the integrated results are returned to improve the efficiency and convenience of users in finding related content.

Images