Survey on Decision Tree Algorithms for Classification

Abstract

Decision Tree (DT) algorithms are among the most widely used supervised learning techniques in data mining and machine learning due to their high classification accuracy and interpretability. This survey presents a comprehensive review of decision tree methodologies, including classical algorithms such as ID3, C4.5, CART, and CHAID, along with recent advancements in multivariate induction, optimal decision trees, and ensemble-based approaches. The survey discusses key components of decision tree construction, including splitting criteria, pruning techniques, and performance evaluation. It also examines the applications of decision tree classifiers in diverse domains such as healthcare, finance, astronomy, remote sensing, and education. Furthermore, the study highlights major challenges, including overfitting, instability, attribute selection bias, privacy concerns, and the impact of data quality on classification performance. Recent developments in cost-sensitive learning, federated learning, and hybrid machine learning models are also reviewed. Overall, this survey provides an overview of the evolution, current research trends, limitations, and future directions of decision tree algorithms for classification in complex and large-scale data environments.