Journal de biométrie et biostatistique

In this study to identify the disease patterns using statistical methods on data of schedule castes of Patna, Vaishali and Nalanda districts of Bihar. Using model based clustering technique; the study is designed to determine the patterns and hidden relationships in dataset. Clustering is a valuable exploratory tool for data analysis that extracts information from a data set and transforms it into an intelligible structure for further applications. The objective of this study to provide profiling of patients, determine dominant disease and dominant month segment. In this regard, clustering is used to profile patients according to their month attended in OPD. The Bayesian Information Criterion (BIC) used to find out the optimum numbers of clusters in a dataset. Using this, a number of clusters are formed on the basis of type of disease acquired by patients, demographic socioeconomic and other characteristics beside that the patients are divided into several clusters based on the diseases they have.