Personalized Recommendation Strategy And Film Recommendation System Design Based On Web Log Mining

With the rapid development of network technology,the Internet has become a huge data carrier for people to enjoy the convenience of "information rich",but it also leads to the problem of "information overload".In this case,the personalized recommendation system provides an effective way for people to locate the data information quickly and effectively.Combined with the development demand of big data era,the method of data mining is utilized to collect the user's interest in the attributed of the items and adjust the user similarity calculation,which improves the recommendation quality of the system.The main contributions of the thesis are listed as follows:1.This paper first designs a personalized recommendation strategy based on Web log mining.Through the collection and processing of user access to the web log data,mining users of the characteristics of the property characteristics of interest.In the calculation of similar users,we design a calculation method based on the similarity degree of the attribute of the item,and integrate the similarity degree of the item attribute with the user rating similarity,and set a balance factor to measure the proportion between them.Similar user sets.Through similar users to the target user did not buy the items to predict the score,the high rating of the items recommended to the target user.Finally,the performance evaluation of the personalized recommendation strategy is compared with the existing solutions before and after the algorithm improvement.The experiment shows that the improved algorithm can provide accurate personalized recommendation service under the condition that the user data is very sparse Compared to other algorithms,the recommended quality is improved.2.Design and implementation of personalized film recommendation system which mainly includes four modules: log collection,data processing,online recommendation and user operation module with the data mining and recommendation algorithm model.The data collection module is mainly responsible for real-time monitoring server log files and monitor the user access data uploaded to the data processing module.The data processing module preprocesses the original log data and excavates the user's interest preference and analyzes the similarity between the users.The recommendation module calculates the neighbor users based on the results calculated by the data processing module and predicts the user's rating and generates recommendations.Among them,the user operation module is the interaction between the user and the system functions including the user registration module,the user login module,the movie details module,the movie search module and the movie score module.3.First,simulate the real user to operate the system for function test through the system function completeness test,the feasibility of the normal operation test and the correctness of the results returned test.Then,the system performance is tested through the ab performance test tool to detect the system in the case of high concurrent compression capacity.