Performance Analysis of Traditional ML Algorithms in High-Dimensional Data with Fuzzy Feature Selection

Abstract

This study investigates the impact of fuzzy feature selection on the performance of traditional machine learning algorithms in high-dimensional data scenarios. We evaluated several algorithms, including Logistic Regression, Decision Trees, Support Vector Machines (SVM), Random Forests, and K-Nearest Neighbors (KNN), using both standard and fuzzy feature selection methods. The results reveal a substantial improvement in predictive performance with the application of fuzzy feature selection. Specifically, accuracy, precision, recall, and F1-score metrics showed notable enhancements across all algorithms, with the most significant gains observed in SVM and Random Forests. The findings suggest that fuzzy feature selection effectively addresses the challenges associated with high-dimensional data by reducing dimensionality and improving signal quality, leading to more robust and accurate models.