Machine Learning and Big Data

8Performance of Supervised Learning Algorithms on Multi-Variate Datasets

Asif Iqbal Hajamydeen* and Rabab Alayham Abbas Helmi

Faculty of Information Sciences & Engineering, Management & Science University, Shah Alam, Selangor, Malaysia

Abstract

Supervised Machine Learning (SML) algorithms stands on the principle of generating theories on the existing data instances to make predictions on the upcoming data instances. Typically, a supervised learning algorithm is provided with a set of labelled instances from which the algorithm generates a model to categorize/ predict future instances. Supervised learning algorithms are used in multidisciplinary research due to its capability in predicting and classifying data accurately, provided the algorithms were sufficiently trained. Therefore, this chapter concentrates on evaluating the performance of supervised algorithms namely Support Vector Machine, Naïve Bayes, Bayesian Network, K-Nearest Neighbour, Hidden Markov Models and Neural Networks. Multi-variate datasets were tested using the aforesaid algorithms to substantiate the suitability of an algorithm and its performance in terms of training time and accuracy with diverse datasets.

Keywords: Machine learning, supervised algorithm, classification, labelled data, accuracy, multi-variate datasets, training time

8.1 Introduction

The initiative for machine learning starts from the straightforward incompetence of the human to learn, comprehend and evaluate large number of features and ...