This paper surveys the methodology of decision tree classifiers (DTCs), discussing their potential and problems, design, and related issues. DTCs are effective in various applications such as radar signal classification, character recognition, remote sensing, medical diagnosis, and speech recognition. They are valuable for breaking down complex decision-making into simpler decisions, making solutions easier to interpret. The paper reviews current methods for DTC design, including tree structure design, feature selection, and decision rules. It also addresses issues like incremental tree design, tree generalization, missing values, robustness, and the relationship between DTCs and neural networks. The paper emphasizes the importance of balancing accuracy and efficiency in DTC design. It discusses various approaches to tree structure design, including bottom-up, top-down, hybrid, and growing-pruning methods. The paper also explores entropy reduction and information-theoretic approaches for improving DTC performance. It highlights the challenges of DTCs, such as overlap in class labels and error accumulation in large trees. The paper concludes that while DTCs have significant potential, their design requires careful consideration of accuracy, efficiency, and robustness. It also suggests that future research should focus on improving DTC performance through efficient heuristics and better handling of complex classification tasks.This paper surveys the methodology of decision tree classifiers (DTCs), discussing their potential and problems, design, and related issues. DTCs are effective in various applications such as radar signal classification, character recognition, remote sensing, medical diagnosis, and speech recognition. They are valuable for breaking down complex decision-making into simpler decisions, making solutions easier to interpret. The paper reviews current methods for DTC design, including tree structure design, feature selection, and decision rules. It also addresses issues like incremental tree design, tree generalization, missing values, robustness, and the relationship between DTCs and neural networks. The paper emphasizes the importance of balancing accuracy and efficiency in DTC design. It discusses various approaches to tree structure design, including bottom-up, top-down, hybrid, and growing-pruning methods. The paper also explores entropy reduction and information-theoretic approaches for improving DTC performance. It highlights the challenges of DTCs, such as overlap in class labels and error accumulation in large trees. The paper concludes that while DTCs have significant potential, their design requires careful consideration of accuracy, efficiency, and robustness. It also suggests that future research should focus on improving DTC performance through efficient heuristics and better handling of complex classification tasks.