抽象的な

Prognostic classification of tumor cells using an unsupervised model

R Sathya Bama Krishna, M Aramudhan

Our human body comprises of trillions and trillions of cells. Cell acts as the basic building block for all living beings. DNA present inside the nucleus of every cell carries genes. Abnormal mutations occurring in these genes are the prime reason for cancerous cell development. A cluster of cancer cells are called tumor. Breast Cancer acts as one of the primitive cause of cancer deaths among women both in developed and developing countries. One of the prominent techniques of data mining is clustering concept, in the field of biology we discuss it as human genetic clustering. The ground rule of any industrial application development lies in appropriate data collection. In this paper a novel clustering technique using an unsupervised data model is proposed. A revised structure based fuzzy C means soft clustering algorithm is proposed and applied over Wisconsin dataset. This algorithm applies a novel strategy in choosing the initial cluster centers. The experiments are carried out over Breast Cancer Wisconsin Data set retrieved from UCI Learning Repository and the work is compared with other existing techniques.