A STUDY OF MACHINE LEARNING ALGORITHMS IN PREDICTIVE DATA ANALYTICS FOR REAL-WORLD DECISION-MAKING SYSTEMS

Abstract

Predictive data analytics has emerged as a critical component of intelligent decision-making systems; however, many organizations still face challenges in achieving accurate predictions from large and complex datasets. This study investigates the effectiveness of selected machine learning algorithms in improving predictive performance for real-world decision-making applications. The research focuses on the comparative implementation of Decision Trees, Support Vector Machines (SVM), Random Forest, and Artificial Neural Networks (ANN) using publicly available datasets from healthcare, finance, and business domains. The datasets were preprocessed through normalization, missing-value handling, and feature selection techniques to improve model efficiency and reliability. The experimental analysis was conducted using Python-based machine learning libraries, including Scikit-learn and TensorFlow. Model performance was evaluated through measurable metrics such as accuracy, precision, recall, F1-score, and prediction error rates. The findings demonstrate that ensemble and deep learning models achieved higher predictive accuracy compared to traditional machine learning approaches, particularly in large-scale and high-dimensional datasets. Random Forest produced an average prediction accuracy of 91.4%, while Artificial Neural Networks achieved 94.2% accuracy in complex classification tasks. In healthcare datasets, the proposed framework improved disease prediction reliability, whereas in financial datasets it enhanced fraud detection and risk assessment capabilities. The novelty of this research lies in its comparative multi-domain evaluation of machine learning algorithms within unified predictive analytics frameworks for real-world decision-making systems. The study further highlights the impact of data quality, algorithm selection, and feature engineering on predictive outcomes. The results suggest that machine learning-based predictive analytics can significantly enhance automated decision-making, operational efficiency, and strategic planning across multiple industries while addressing challenges related to scalability and data-driven intelligence.

Keywords: Machine Learning, Predictive Analytics, Decision Trees, Artificial Neural Networks, Support Vector Machines, Random Forest, Real-world Decision Systems, Data Mining, Artificial Intelligence, Predictive Modeling