This paper discusses about detecting frauds in healthcare insurance from large databases using clustering technique. The paper applies two clustering methods SAS EM and CLUTO. The methods include defining data and the preprocessing steps and also data transformation. The last part of the paper compares the two clustering