Effective Learning and Classification using Random Forest Algorithm

Random Forest is a supervised machine learning algorithm. In Data Mining domain, machine learning algorithms are extensively used to analyze data, and generate predictions based on this data. Being an ensemble algorithm, Random Forest generates multiple decision trees as base classifiers and applies majority voting to combine the outcomes of the base trees. Strength of individual decision trees and correlation among the base trees are key issues which decide generalization error of Random Forest classifiers. Based on accuracy measure, Random Forest classifiers are at par with existing ensemble techniques like bagging and boosting. In this research work an attempt is made to improve performance of Random Forest classifiers in terms of accuracy, and time required for learning and classification. To achieve this, five new approaches are proposed. The empirical analysis and outcomes of experiments carried out in this research work lead to effective learning and classification using Random Forest algorithm.