Clustering

Identification and separation of rare or unusual cases from the normal ones has recently gained a lot of popularity.It helps to understand and analyze the data in an efficient and effective manner.It applications in image analysis, web use data analysis, bio informatics, statistics makes it more convenient and acceptable to be use, especially in modern research.It has been observed that on average about 15% - 20% cases shows some unusual or uncommon behavior.Which is a big percentage and contains very important and interesting results that cannot be just discarded from normal data. Several techniques have been widely used to identify rare cases. The purpose of this research work is to identify the most appropriate methodology for the identification of rare cases. Within this research, all the popular technologies are studied in detail and after learning from them, a new methodology has been proposed which gives better results on its application on real data, that has been collected from various hospitals situated in various countries all around the world. So this work consist of research work backed by its practical applications.