Research Of Personalized News Recommender For Mobile Network

With the rapid development of Internet technology and the gradual integration of the mobile communication network and internet network, more and more users access the network services and information by using mobile devices (such as smart phones, PAD, Tablet PC, etc.). However, when users enjoy a variety of mass information, they also face a serious problem of information overload. In addition, due to the limitations of mobile terminal’s processing power, network bandwidth, screen size, the traditional navigation and retrieval technology are unable to meet the personalized needs of the users.Mobile recommendation system provides people with a new personalized service model. It collects and forecasts the mobile user’s browsing behavior and preferences, then pushes information for the users to meet their individual needs by using personalized recommendation analysis technologies. Therefore, the mobile recommendation system can overcome information overload and obtain a lots of people’s attentions.By analyzing the characteristics of mobile Internet different with the traditional Internet, we designed a personalized recommendation algorithm for mobile Internet, and implemented a mobile healthy news recommendation system.The main contributions of this paper are as follows:1. To save in the mobile network traffic and improve the transmission efficiency, a document summarization algorithm based on LDA is proposed. Firstly, we calculate the similarity of topics probability distribution between document and sentence as a new feature. Then, traditional summarization features such as position of sentence in a text and topic similarity are considered. Finally, summary are generated by selecting the sentences with the highest scores. Experimental results show that the performance of our method is outperforms the traditional methods when the combined features join into the LDA Model.2. This paper proposes a hierarchical collaborative filtering recommendation algorithm based on the social tagging and the temporal interesting modeling. Firstly, the reduced user-book-tag tensor model is adjusted by the interest transferring curves, which fitted by the temporal tagging behavior of each tags. Then, the candidate social tags are extracted from the social community by the rebuild User-Tag matrix C. After constructing the user model and the item model by matrix factorization, we use the naive Bayes classifier to recommend top-n books.3. Using the proposed algorithm and Android developing technology, a personalized healthy news recommendation system for mobile Internet is designed and implemented.