Research On Personalized News Recommendation Algorithm Based On Micro Blogging

With the development of Internet technology, the way people get their news is also undergoing a significant change from the traditional newspapers and magazines to news aggregation site, such as domestic services Netease news, Sina news, foreign services Google News, Yahoo News. We are entering the era of information overload from the lack of information era. For the news website, the most important aspect is how to effectively to get the user’s profile by analyzing the user’s interests, then recommend news to users. In recent years, with the rise of micro blogging and other social networks, many scholars have tried to build user’s profile by analyzing the user’s micro blogging contents and social behavior. It has become a new hotspot for studying user’s interest based on a micro blogging.In this paper, We research and conduct the personalized news recommendation algorithm based on the micro blogging. The main contents are as follows:For the words limit of micro blog, We utilized POS tagging to enrich the content of micro blog to express the user’s interest better; As news’s attributes are not complete and improving user experience. We have designed a combined text classifier to classify news. And We also give an intelligent algorithm to automatically generate a summary of the news. To conquer the cold-start problem of the recommend system. Inspired with Chinese Restaurant Process, We provide a word vector based recommend algorithm. To better understand the relationships between user and news. We model them using tensor, We put forward a recommend algorithm based on tensor factorization; By experimental comparison, the proposed recommendation algorithm is better than traditional keyword-based content recommendation algorithms. Even though the number of a user’s micro is less, the effect of the algorithm is also very good, the algorithm has a better fault tolerance.This paper also presents a complete design solution personalized news recommendation system based on micro blogging, and elaborated on the system design and implementation. Through the monitoring and analysis of system, We confirmed the effectiveness of our algorithm and availability of the system.