Optimized Predictive Technique for Intelligent Classification (OPTIC): A Decision Tree Ensemble Approach for Neurodegenerative Disease Detection via Gait Pattern Analysis

Abstract

Purpose of the Study: Neurodegenerative diseases such as Parkinson’s, Alzheimer’s, and Huntington’s progressively impair motor and cognitive functions, making early detection essential for timely intervention and improved quality of life. This study introduces the Optimized Predictive Technique for Intelligent Classification (OPTIC), a novel decision tree ensemble framework, designed to enhance the accuracy and robustness of gait-based neurodegenerative disease detection. Methodology: OPTIC employs a modular ensemble architecture that partitions the feature space into subsets, trains separate decision tree modules, and integrates their outputs through collaborative feedback and dynamic weighting based on classifier complexity and performance. The model was evaluated on a publicly available gait dataset using multiple validation strategies, including holdout, stratified k-fold, and Monte Carlo cross-validation. Key Results: OPTIC consistently outperformed traditional decision trees and standard ensemble methods. Under stratified k-fold validation, it achieved 78.98% accuracy in multiclass classification, 91.48% accuracy in one-vs-rest settings, and up to 99.13% accuracy in pairwise disease-versus-control classification. Performance improvements were consistent across validation schemes, with OPTIC demonstrating superior precision and recall in all cases. Conclusion: The OPTIC framework effectively addresses limitations of conventional decision trees, such as overfitting and instability, through modular specialization and collaborative learning. Its high accuracy and robustness suggest strong potential as a reliable tool for early detection of neurodegenerative diseases, with practical applicability in clinical decision-making and patient monitoring.