A Comprehensive Study of Machine Learning Techniques for Predictive Data Analytics

Machine learning (ML) is a relatively new discipline that allows computers to learn based on their data and decide independently without the necessity to be programmed. Since other sectors of the economy such as medical, financial, educational, and social media have increasingly generated more and more digital data, there has been an increased desire to apply Predictive Analytics to analyze the data. It can be done by machine learning techniques, which this research shows in a comprehensive manner of how to use machine learning to predict analytics, with the three key approaches being supervised, unsupervised and ensemble techniques. The benchmark dataset is publicly available and is used to evaluate and compare the effectiveness of various algorithms on several pre-established assessment metrics, including that of accuracy, precision, recall, F1-score, and RMSE. The article provides an analytical study on the pros and cons of both machine learning methods. These findings clearly confirm that, most of the times, the performance of ensemble learning techniques is superior to that of the conventional machine learning models. The results provide significant information to the researchers and practitioners who are interested in implementing machine learning methods to real-world forecasting problems.