A Comparative Analysis of Filter and Optimization Methods for Feature Selection

Selecting the most important features or attributes from a dataset is a crucial part of machine learning and data analysis. This paper explores the effectiveness of feature selection (FS) methods, comparing filter based techniques such as Pearson Correlation Coefficient (PCC), Spearman Rank Correlation Coefficient (SCC), and Mutual Information (MI) with optimization techniques including Genetic Algorithm (GA), Grey Wolf Optimizer (GWO), Cuckoo Search (CS), and Particle Swarm Optimization (PSO). The evaluation involves diverse datasets with varying sizes and complexities. The results show that the choice of FS method significantly impacts accuracy. Filter methods perform better with smaller datasets, while optimization algorithms show adaptability across small and high dimensional datasets. This paper shows that the effectiveness of feature selection is closely tied to the specific dataset being analyzed.