Information retrieval method and system oriented to user-defined user intention model

Abstract

The invention is applicable to the technical field of personalized information retrieval, and provides an information retrieval method oriented to a user-defined user intention model. The method comprises the following steps: S1, acquiring a query log of a user, and calculating the relevancy between words by using a local context analysis method to obtain a query expansion word set; S2, according to a query statement input by a user, searching a corresponding recommended document in combination with the query expansion word set; S3, sorting the recommended documents by utilizing the similarity scores and the importance scores; S4, finding out words with relatively high similarity scores and importance scores in each recommended document, carrying out label annotation on the query document, and feeding back to a user; and S5, calculating the satisfaction degree of the user to the query result based on a positive and negative feedback query optimization method. The method is used for solving the problems that: in the prior art, query conditions input by a user often belong to short texts, fuzziness, mistakes, omissions, non-standardization and the like exist, user intentions are difficult to accurately understand from the query conditions, and feedback programs of the user for query results are lacked.

Description

technical field [0001] The invention belongs to the technical field of personalized information retrieval, and in particular relates to an information retrieval method and system oriented to a user-defined user intention model. Background technique [0002] Personalized information retrieval obtains user preferences through long-term observation of user browsing behavior, and it can track information related to the use of users to help improve retrieval performance. The purpose of tracking user-related information is to obtain the real needs of users and use the captured relevant information to optimize query results. [0003] For the scenario of personalized retrieval, traditional methods consider the respective content of query and information when matching query and information, and synthesize the correlation between information and query to give the final retrieval and ranking results. However, since users' query sentences are often short texts, there are problems such ...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide an information retrieval method and system oriented to a user-defined user intention model, which is used to solve the problem that the query conditions input by users in the prior art often belong to short texts, and there are problems such as ambiguity, errors and omissions, irregularities, etc., which are difficult to solve. Accurately understand user intent from it, and lack of user feedback procedures for query results

Images