Certain investigations on multiclass clustering and fuzzy classification of multi labeled data

Abstract

A Supervised and Unsupervised Learning technique of Machine Learning has been used to analyse and turn data into useful knowledge. Clustering is an unsupervised leaning technique; Classification and prediction are some of the widely used techniques in Supervised Learning. The above mentioned techniques are used to extract the useful knowledge in binary data, text data, spatial and temporal data set. Most of the algorithms for unsupervised learning like clustering is used to group the data objects into cluster at a single-concept level or multiple levels. newlineDuring the past two decades, a numerous amount of research applications has been taken up for machine learning tasks. Most of the study made in this area has been applied through supervised and unsupervised learning algorithms which focused to design Multiclass normalized clustering and classification model framework for machine learning techniques. The framework comprises of three phase such as Pre-processing, Multiclass Clustering and Fuzzy Classification Model. Pre-processing task takes the large data set as input and normalizes the given input to transformed or consolidated for appropriate clustering task. An input data has been equipped to perform all learning algorithms. newlineReal time applications of web based data such as regional Blood bank and Electricity consumption data has motivated to propose an optimization algorithm for Multiclass Normalized Clustering and developing classification model to analyze the real time data. A new Enhanced Multiclass Normalized Optimal Cluster algorithm has been introduced to find the optimum cluster size for the given data objects. newline