Design And Implementation Of User Behavior Feature Extraction System Based On News Fields

The news website essentially provides a news information services for users. Users show different concerns for different kind news. How to find users real interest and give recommendations according to their interest put forward new demands for news website.This paper implements a user behavior feature extraction system which is a part of new recommendation system. This system can be divided into five modules:the input module,the user behavior feature extraction module,the recommendation rule generation module,the output module and administration module. The core part is user behavior feature extraction module. The feature extraction algorithmic determines not only the precision of recommendation rule but also the quality of system and the satisfaction of users. The user feature extraction algorithmic in this paper reference from attribute reduction and decision rule extraction of rough set theory.This paper mainly achieved:1) A reduction and feature extraction algorithmic which can calculate the importance of different kind of news according user browse record. And then extract the interest value of different kind of news called user behavior feature. This algorithmic which is the research point of paper is the base of news recommendation rule and the quality of system.2) Recommendation rule algorithmic is to extract news recommendation rule for particular user and it can generate different rules for different user. The news recommendation rule generation part is based on feature extraction part and it is the complement of the system. After function test and integration test this system could extract user particular behavior feature according to user browse behavior and extract news recommendation rule to display news to user. The recommendation result fit user interest and have better conformity.