摘要
研究电信客户分类问题,根据不同类型采用不同策略。针对电信客户实行差异化营销和服务,需对电信客户进行准确分类。传统的k-均值聚类算法是一种重要数据挖掘技方法,存在对初始值敏感和易陷入局部最优的缺陷,导致电信客户分类正确率较低。为了提高电信客户分类的正确率,提出了一种改进k-均值聚类的电信客户分类算法。首先改进k-均值聚类算法通过变异、杂交和选择操作,然后根据分类特征动态地确定初始聚类数k和自适应确定聚类中心,最后采用湖南省某地区客户分类数据进行验证性实验。仿真结果表明,改进k-均值聚类算法很好地解决全局识别寻优问题,提高了客户分类正确率,大幅度减小误差。
In order to implement differentiation of telecommunications customer marketing and customer service strategies,customers need to be accurately classified.K-mean algorithm is an important method for telecom customer classification in the data mining technology,but in actual classification process,k-means clustering algorithm is sensitive to initial value and easy to fall into local optimal.In order to improve the telecom customers＇ classification accuracy,a new clustering algorithm is put forward to improve the classification accuracy.The improved k-means clustering algorithm＇s initial clustering K and clustering center are adaptively determined by hybridization,operator selecting and classification feature.Simulation experiments show that the improved k-means clustering algorithm can solve the problem that traditional k-means clustering algorithm easily trapps into local optimal,improve customer classification accuracy and reduce errors greatly.
出处
《计算机仿真》
CSCD
北大核心
2011年第8期138-140,152,共4页
Computer Simulation
关键词
均值算法
差分演化算法
客户分类
Mean algorithm
Differential evolution algorithm
Categorized customer
