Community question and answer website label recommendation method based on user background

Abstract

The invention discloses a community question and answer website label recommendation method based on a user background, comprehensively considering the to-be-recommended problem text information, theuser background information and the relevance of the text information and the user background information, and modeling a label recommendation problem into a multi-classification prediction problem based on deep learning. The core of the community question and answer website label recommendation method is a deep neural network model PcTagger, and through dynamic modeling of the user background information, the defect that user background static modeling in an existing personalized label recommendation method is difficult to match different recommendation tasks is solved. The model mainly comprises: 1) text feature modeling based on a recurrent neural network and an attention mechanism; 2) user background influence dynamic modeling based on historical question records, and 3) label recommendation by fusing the text features and the user background influence. The experimental results displayed on a real data set show that the community question and answer website label recommendation method based on a user background can significantly improve the prediction precision, compared with an existing same kind label recommendation method.

Description

technical field [0001] The invention relates to a tag recommendation method for a community question-and-answer website based on user background, and belongs to the technical field of text tag recommendation. The invention utilizes the serialized text content and the user background influence extracted from the user's historical behavior, and based on the deep neural network model, improves the recommendation accuracy of the question tags of the community Q&A website, and effectively improves the tag recommendation effect for new questions. Background technique [0002] The development of the Internet has made the exchange and sharing of knowledge convenient and fast. The community Q&A website represented by Stackoverflow provides a platform for technicians to exchange questions and share methods, which greatly improves the efficiency of learning and work. Due to the rapid growth of information, users of these platforms face the problem of information overload to varying de...

AI Technical Summary

Problems solved by technology

However, this method has fixed user personalized information, and cannot accurately reflect the influence of user background in the face of changing content to be recommended.

Images