An Efficient Hierarchical Clustering Technique for Medical Diagnosis Using KNN Classifier

In this research article, an intelligent hierarchal clustering technique for medical diagnosis system has been proposed. Various hierarchical clustering techniques and their variants have been very much explored in the field of machine learning. However, these techniques are deterministic, needn't bother with a determined number of clusters and are stable. But, they are not scalable for high dimensional data set due to their non-linear correlations. In this paper, a new approach is proposed for medical data classification based on hierarchical clustering. The proposed technique has the following features (i) In each cycle, rather than ascertaining the centroids for new clusters, new centroids are assessed from centroids in past cycle; and (iii) In every run, rather than combining just a single match of items, various sets are converged in the meantime.