Greedy Binary Search and Feature Subset Selection

In this paper, we propose new Greedy Binary Search methods, Greedy Binary Search (GBS) and Greedy Binary Search with Parameter(PGBS), to be used for feature selection problem. We describe the traditional algorithms including SFS (SBS), GSFS (GSBS), PTA (l, s) (GPTA (l, s)) and SFFS (SBFS). We also explain our proposed methods and compare them with the time complexity. Through a variety of experiments and analysis, it is found that our methods outperform traditional algorithms in terms of the number of features and the correct classification rate.