摘要
从企业产品数据管理需求的现状出发,针对数据挖掘中经典的k-means算法中存在的不足并考虑到PDM系统中存在的数据量相当大、数据类型复杂等现实问题,参考采用多次取样一次聚类寻找最优解的改进算法,并通过模拟系统实验验证该算法的稳定性及效果有显著的改善。
From the current situation of enterprise product data management needs,aiming at shortage of data mining in the classic k-means algorithm that exist and considering the amount of data that exist in the PDM system is quite large,complex data types and other practical problems,a reference to the use of multiple sampling improved clustering algorithm to find the optimal solution by simulating experiments verify the stability of the system and the algorithm is significantly improved.
出处
《计算机与数字工程》
2016年第11期2213-2217,2256,共6页
Computer & Digital Engineering
基金
无锡市教育科学"十二五"规划立项课题<非标企业PDM管理系统的实践与研究>(编号:J/D/2014/025)资助
