Abstract: Clustering mixed dataset is a very common approach in day to day life. Several algorithms exists for clustering heterogeneous datasets with accuracy. But most of the algorithms are inefficient and unable to cluster only one type of records. Therefore we have designed an algorithm which is efficient as like k means and can also handle mix type and only one type of data in dataset. Our algorithm has been tested on standard dataset to see the performance.

Keywords: Heterogeneous, Fuzzy, Cluster, Centroid, Partition.