A clustering algorithm for categorical data with combining measures

Clustering is one of the main techniques in data mining. Clustering is a process that classifies data set into groups. In clustering, the data in a cluster are the closest to each other and the data in two different clusters have the most difference. Clustering algorithms are divided into two categories according to the type of data: Clustering algorithms for numerical data and clustering algorithms for categorical data. The clustering algorithms for categorical data are more important than clustering algorithms for numerical data because of the nature and application of these data. Each of these algorithms uses different similarity measures according to the type of data (numeric or categorical). In this paper, a new clustering method is proposed for clustering by combining Overlay and Jaccard similarity measures on a hierarchical algorithm for categorical data. Overlay measure represents similarities between the data as one and zero which caused the loss of some information. Jaccard measure If used alone to measure the similarity between data set, most clusters are selected in the particular area of data collection. So in this paper a combination of the two measures are used. Experimental results show that the proposed method improves the results of clustering. Resulted improvemen is 10% on any evaluation factor in average.