Analysis Of User Interest Based On Mobile Internet And Its Application

With the rapid development of information technology and mobile Internet,mobile intelligent terminal has become increasingly popular,and bring us into the era of information explosion,people receive countless spam every day.Especially the portability and the instantaneity of mobile Internet makes people to get a lot of information anytime and anywhere.Users are eager to get useful information from vast amounts of information.This is a new requirement for Internet content providers to provide services with precision and specificity.The key to solve this problem is accurately grasping the needs of users,and providing accurate and personalized information push service according to their interest.Therefore analyze and study how to get the user interest is of great significance.This dissertation studies the calculation of user interest based on the WEB browsing behavior analysis and modeling method,improves the model performance by introducing situational factors in the traditional hierarchical vector space interest model,improves the traditional K-means clustering algorithm and the Fuzzy Cmeans clustering algorithm and uses the improved algorithm to analyze the users' interests,and obtains a clear improvement in the accuracy of the acquisition and analysis of user interest.Finally this dissertation designed a user interest analysis system based on JFinal open source framework.This main research content of this dissertation is:(1)the user interest calculation and modeling.Through the analysis of user browsing activity in the process of network access,find and calculate the user's interest,and propose the calculation method.(2)improve the traditional interest vector space model.By introducing situational factors to the traditional hierarchical vector space interest model,it get more versatile and adaptive..(3)improved the traditional K-means algorithm and the traditional Fuzzy Cmeans algorithm.On the basis of the traditional K-means algorithm,the weighted Euclidean distance are introduced to the initial clustering center which is determined based on the density,improved the algorithm accuracy.In allusion to deficiency of the traditional Fuzzy C-means algorithm in easy to fall into local minimum solution,by introducing the theory of natural selection and evolution of genetic algorithm,the new algorithm improves the clustering effect significantly.(4)the user interest analysis system.Based on JFinal open source framework developed a user interest analysis system,can detailed analysis of the user's browsing behavior to find out the user potential interest,and according to the user's interests to realize recommendation,has good applied significance.