An Effort to Compare the Clustering Technique on Different Data Set Based On Distance Measure Function in the Domain of Data Mining

[摘要] Abstract. Clustering divides a database into different groups to find groups that are very different from each otherand whose members are very similar to each other. There are many clustering approaches all based on the principle of  maximizing  the similarity  between  objects  in  a  same  class  ( intra-class  similarity )  and minimizing the similarity between objects of different classes ( inter-class similarity ). This difference has been calculated based on the some distance measure function. It has been observed that most of the authors used the clustering techniques to select the optimal cluster for the particular data set. But they did not  made the comparison on selection of the optimal cluster based on the distance measure function. In this paper an effort has been to select the optimal cluster based on difference distance measure function of cluster.   On distance measure know as Bit equal has also been proposed and its performance has been compared with other existing distance measure function. The K-means algorithm has been used on all the data set to select the optimal clusters.

[学科分类] 建筑学

[关键词] Cluster;K Means;Hierarchical;Euclidean;Hammingdistance;Bitequal and Mahalanobisdistance function.