Search engine user satisfaction evaluation method based on long-short time memory network

Abstract

The invention discloses a search engine user satisfaction evaluation method based on a long-short time memory network. The method comprises the following steps of firstly, extracting a search behaviorsequence from a search engine log, performing the data enhancement on the search behavior sequence by using a data enhancement strategy based on the time interval perturbation, and introducing a virtual staying behavior to represent a time interval between the continuous behaviors; secondly, learning the feature representation of the search behavior sequence by utilizing the long-short time memory network, and establishing a search engine user satisfaction classifier; and finally, identifying the user satisfaction label of the given sample by using the constructed user satisfaction classifier. According to the method, the search engine user satisfaction is more effectively evaluated through the feature representation of the end-to-end learning search behavior sequence.

Description

technical field [0001] The invention relates to the field of Internet information technology, in particular to a search engine user satisfaction evaluation method based on a long short-term memory network. Background technique [0002] Search engines are one of the main ways for users to obtain the required resource information from the massive Internet data. As users have higher and higher requirements for efficient and convenient access to information resources, search engines need to continuously optimize the system to provide users with better search services. Therefore, how to effectively evaluate the quality of search engines has become the focus of research and industry. [0003] Search engine logs record a series of behavioral information that occurs during the interaction between users and search engines, and search satisfaction has a strong correlation with users' search behaviors. Therefore, researchers usually use the search behavior sequences extracted from se...

AI Technical Summary

Problems solved by technology

[0005] In addition, compared with non-deep learning methods, deep learning methods require more labeled data. Due to the expensive and time-consuming process of collecting labeled data, the trained model may lead to overfitting due to insufficient training data.

Images