基于聚类分析的数据挖掘方法研究

Study on Data Mining Based on Clustering Analysis

针对FCM算法的聚类效果易受其初始中心位置影响和易陷入局部最优的缺点,将灰狼优化算法和FCM结合,提出一种基于GWO优化FCM的聚类分析方法。以KDD CUP99数据集为研究对象,研究结果表明,与PSO、GA和SA算法相比较,GWO算法聚类分析的准确率和误判率更低,具有更快的收敛速度,效果更优,从而为数据聚类分析提供新的方法和途径。

Aiming to the clustering effect on FCM algorithm is easily affected by the initial location and fell into local optima,the gray optimization algorithm is applied to improving FCM to put forward a method of clustering analysis GWO_FCM.The KDD CUP99 data set as the research object, the research results showed that GWO algorithm of clustering had a accuracy and lower error rate, comparing with PSO, GA and SA algorithm, it was with faster convergence speed and better effect so as to provide a new approach for data clustering analysis.