Hybrid Clustering and Classification

In data mining, C-Means clustering is well known for its efficiency proved good for large data sets. The aim of every clustering algorithm is to group the similar data items while ungroup the dissimilar items. C-Means clustering algorithm has the opposite principle as fuzzy clustering algorithm has i.e. in C-Means every point has belonging to clusters while in fuzzy clustering, they belong to only one cluster.Clustering is a supervised learning algorithm. Clustering dispersion called entropy factor is the disorderness that occur after the clustering process. Less entropy leads to good clustering. Clustering with C-mean results in unlabeled data. I present a clustering algorithm called C-Means. Then unlabelled data is matched through neural classifier. Neural Network is the classification function to distinguish between members of the two classes in the training data. For classification we use Neural Network as they can recognize the patterns. The whole work is taken place in the Mat lab 7.10 environment in which entropy is taken as the main parameter for performance.