一种基于基因表达式编程的串行聚类算法并行化研究  

Research on GEP-based Cluster Algorithm for Serial Program to be Parallelized

作　　者：蔡宏果[1,2] 元昌安[1] 

机构地区：[1]广西师范学院科学计算与智能信息处理广西高校重点实验室,南宁530023 [2]广西教育学院培训学院,南宁530023

出　　处：《中南民族大学学报（自然科学版）》2017年第4期112-115,共4页Journal of South-Central University for Nationalities：Natural Science Edition

基　　金：国家自然科学基金资助项目(61262028)

摘　　要：为进一步解决基于用户的协作过滤技术的扩展性问题,利用基因表达式编程(GEP)的并行性优势,与已有的串行聚类DBSCAN算法进行融合,使得串行程序并行化,提出了一种GEP-DBSCAN协作过滤聚类算法来寻找最近邻居,改进基于密度的协作过滤方法,实验证明了算法的有效性以及提高了时间效率.To address the expansibility problem of collaborative filtering technology based on users,the GEP-DBSCAN algorithm for collaborative filtering clustering was proposed.It is key to fuse the parallelism of gene expression programming and the advantages of the DBSCAN algorithm.The new algorithm makes serial programs parallelized and can be used to find the nearest neighbors.It improves collaborative filtering method based on the density.The experimental results show that the GEP-DBSCAN algorithm is effective and can increase time efficiency.

关 键 词：聚类算法 基因表达式编程 协作过滤 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]