Chameleon聚类算法的Weka实现

为了克服Weka系统在聚类算法方面的不足，在Weka的开源环境下进行了二次开发，扩充了聚类算法。介绍了Chameleon算法的基本原理和构建步骤，给出算法的具体描述，并将Chameleon算法嵌入Weka平台，充分利用了其中的类和可视化功能。对实现的系统进行了实验和测试，并对嵌入的算法和原有聚类算法k—means进行了对比分析。实验结果表明，Chameleon算法可获得更好的聚类效果。

Study of Chameleon Clustering Algorithm and Implementation in Weka

To overcome the Weka＇s weakness in clustering algorithms, the clustering algorithm was developed based on the Weka open source environment. An introduction is made on the basic principle and construction steps of the Chameleon algorithm which was described concretely and embedded into Weka platform whose inner class and visualization function were exploited adequately. The newly built Weka system was tested by comparing the embedded and the already existed K-means algorithm, The result shows that Chameleon algorithm acquires better clustering effect.