面向社交网络密集图数据存储的缓存置换算法研究.

In order to alleviate the problems of frequent I/O of storage and space waste caused by dense graph data generated by hot topics in social networks, it propose a cache replacement algorithm based on topic heat evolution acceleration (THEA-CR) according to the evolution and update law of topic generation and death. Firstly, this algorithm divided the social network data into some topic clusters to identify anchor targets. Secondly, this paper calculated the topic heat evolution acceleration to evaluate and judge the priority of hot data. Finally, this paper designed a dual queue cache replacement strategy to replace and update the cache space according to topic concern and access frequency. Massive comparative experiments verify the feasibility and effectiveness of the proposed algorithm in Sina Weibo dataset with the baselines cache replacement algorithms. The results show that the proposed THEA-CR can improve the cache hit rate by about 31.4% on average and shorten the query response time by about 27.1% in different query operations of social network dense graph data. [ABSTRACT FROM AUTHOR]