Evaluation of Principal Components Analysis (Pca) and Data Clustering Techniques (Dct) on Medical Data

The present study investigates the performance analysis of PCA filters and six clustering algorithms on the medical data (Hepatitis) which happens to be multidimensional and of high dimension with complexities much more than the conventional data. By Clustering process data reduction is achieved in order to obtain an efficient processing time to mitigate a curse of dimensionality. Usually, in medical diagnosis, the chief guiding symptoms (rubrics) coupled with the clinical tests help in accurate diagnosis of the diseases/disorders. Hence, the primary factors have maximum impact/influence on the detection of the specific disorders. Therefore, the present study is undertaken and the results predict that farthestfirst clustering algorithm happens to be the best clustering algorithm without PCA filter in general, while cobweb clustering algorithm could be preferred with PCA filter in some other medical datasets