Application Of Improved Algorithms Of Clustering To Recommendation Of Long Tail Text In Online Self-publication System

With the developing of network technology, the Internet plays a growingly important role in many aspects of people's living and studying. Then the personalized information service is much more attractive because of the plenty of resources on the Internet. Meanwhile, people's aim for getting information has been ranged from checking-all to checking-exactly and the new goal of the information service in the present network environment has become to satisfy the special needs of particular men. Recommendation System (RS) is born in this environment. RS, using data mining technology and AI technology, provides intelligent and personalized service. It could recommend pages or products to the customers, and could lead customers to focus on the information that they want.The general public's entering in web content producing as the development of technology has supported self-publication's success in US, which provides a great potential point of growth in network market. Theory of Long Tail refuted the classical economic theory and figured out the same power between the main markets and accumulated small markets. To deal with the self-publication system's huge text scale and speedy increasing, a right recommendation technology is helpful to realize the market value of "Long Tail" books. For this issue, Directed Graph Clustering Algorithm is presented. Regarding clustering as the process of objects identifying, key words' election depends on how much information they transfer in context. The experimental results comparing the K-Means algorithm prove that this clustering algorithm based on the directed graph is self-adaptive and effective.The core of RS is the recommendation method, which is also the core of this article. To improve the efficiency of "Long Tail" texts'recommedation, a floating threshold and a sharing pool are set. It designed a whole system which has three processes:date washing, date mining and recommendation, and proved the feasible and effectual of this system based on the theory of Long Tail logically and the results of the imitating experiments actually.