Cache Replacement Algorithms Based On Data Mining

With the development of Internet technology, network information and the number of Internet users has exploded, resulting in heavy network traffic and network congestion and so on. Currently, the cache is the primary measures to address these problems. Cache replacement algorithm has been the main research to improve cache system performance. Most of the current cache replacement algorithm is analyzed from the perspective of the cache object, ignoring the real Internet users.Therefore,from the users’ point of view, we use the method of statistical analysis of data mining to analyse the real video log data for exploring the characteristics of access behavior of users,and use these characteristics to design new cache replacement algorithm.The main research and contributions are as follows:First, we use the method of statistical analysis of data mining to analyse the real video log data for exploring the characteristics of access behavior of users. By researching the access motivation of users, relationship between the activity of users and the importance of video file objects,we found the following phenomenon:the video file object that has low average activity of user will be accessed frequently in the future. This discovery provides a reference value for designing cache replacement algorithms, recommending the files and classifying different users.Then, we propose a caching replacement strategy based on user activity. We take average activity of user as value function which is as measuring video file objects in the cache. When the video file objects of cache need to be replaced, replacing those video file with higher average activity of user.Simulation results show that the algorithm can achieve better hit rate and byte hit ratio in the case of large cache capacity.Finally, we propose a caching replacement strategy based on PageRank algorithm.The algorithm treat user and video file objects as nodes in the network, treat the access relational between user and video file objects as the network connection, and then use PageRank algorithm to weight users and video file objects. According to the weight value of the video file objects,we design the cache replacement algorithm, When the video file objects of cache need to be replaced, replacing those video file with lower weight value.Simulation results show that the algorithm can achieve better hit rate and byte hit ratio.