A clustering hybrid algorithm for smart datasets susing machine learning

Abstract

In the field of data science, Machine Learning is treated as sub-field which primarily deals with designing of algorithms which have ability to learn from previous information and make future predictions accordingly. In traditional computational world the Machine Learning was generally performed on highly performance servers and machines. The implementation of these concepts on Big Data analytics algorithms has high potential and is still in its early stages. So far as machine learning is concerned, performance measure is an important parameter to evaluate the overall functionality of the algorithms. The data set is a different entity and the measuring of performance on a data which is unseen is also called as test set, and training set is a Data set which is training itself. The Data Mining is extensively using learning algorithms for data analysis and to formulate future predications based on archived data. The research presented provides a step forward to make smart data sets out of training data set by evaluating machine learning algorithm. The research presented a novel hybrid algorithm that attempts to incorporate the feature of similarities in Random Forest machine learning algorithm for improving the classification accuracy and efficiency of working.