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 the GEP- based Cluster algorithm for serial program to be parallelized
投稿时间:2017-06-27  修订日期:2017-06-27
DOI:
中文关键词: 聚类算法  基因表达式编程  协作过滤  
英文关键词: Cluster algorithm Gene Expression Programming collaborative filtering
基金项目:国家自然科学基金:61262028;广西自然科学:2012GXNSFAA053235;通讯作者:*元昌安(1965-),男,教授,博士,研究方向为数据挖掘、智能计算,E-mail: webminning@163.com
作者单位E-mail
蔡宏果 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 webminning@163.com 
元昌安 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 yca@gxtc.edu.cn 
摘要点击次数: 19
全文下载次数: 
中文摘要:
      为进一步解决协作过滤技术的扩展性能问题,依据基于密度原理的聚类思想,利用基因表达式编程(Gene Expression Programming,GEP)的并行性,对已有的串行聚类算法进行改进,使得串行程序并行化,提出了一种基于GEP的协作过滤聚类 (GEP-DBSCAN) 算法来寻找最近邻居,提高基于用户的协作过滤技术中用户聚类的时空效率,为个性化推荐系统服务。实验证明了算法的有效性。
英文摘要:
      According to the theories of the density-based clustering and the parallelism of the GEP (Gene Expression Programming), the GEP-DBSCAN algorithm was proposed for further solving the expansion of the collaborative filtering technology performance problem. The GEP-DBSCAN algorithm improves traditional serial cluster algorithm that makes serial program to be parallelized. The algorithm can reduceStheSspace overhead for the collaborative filtering. It can solve the nearest neighbor problem for services of personalized recommendation system based on users. The results indicate the efficiency of the algorithm.#$NLKeywordsCluster algorithm; Gene Expression Programming; collaborative filtering
View Fulltext   查看/发表评论  下载PDF阅读器
关闭