| Summary: | One of the most threaten disease in the women is Breast Cancer. It may also spread in to remaining parts of the body slowly. So the single drug treatment for curing breast cancer is not efficient and combinational therapy will be better. The combinational therapy involves combing the one or more drugs with specific dosage of each, this will improve the efficacy and it also reduces the single drug dosage for the treatment of Breast Cancer. As multiple drugs are used in combinational therapy, disease development may be delayed, reduces the growth of tumor. In this paper we propose the clustering of breast cancer drugs based on their physico chemical properties. First the values of important attributes for the selected breast cancer drug dataset are computed. By applying Hopkins statistic we can found that the dataset is highly clusterable and after that the elbow method identifies the optimal number of clusters. The PEMFKM clustering algorithm was proposed and applied on the dataset to produces three well defined clusters with clear differentiation. During the cluster validation analysis, variant fuzzy clustering algorithms like FKM, FKM.ENT, FKM.ENT.NOISE, FKM.GK, FKM.GK.ENT, FKM.GK.ENT.NOISE and the PEMFKM are compared. The index values of parameters PC, PE and MPC are better for FKM.GK when compared to the proposed PEMFKM algorithm. The index values of parameters SIL, SIL.F and XB are better for proposed PEMFKM algorithm when compared to FKM.GK algorithm. Thus, we conclude that PEMFKM algorithm will form better clusters for the selected drugs, and the drugs within the same cluster may be combined to form more effective drug and use in the combinational treatment of Breast Cancer treatment.
|
|---|