A Binary Chaotic Transient Search Optimization Algorithm for Enhancing Feature Selection

Real-world data mining problems require feature selection to improve efficiency and accuracy. Due to not considering characteristics of the FS problem itself, traditional mechanisms limit their performance on dealing with high-dimensional FS problems. Focused on it, this paper proposes a novel feature selection algorithm based on binary transient search optimization (TSO) with chaotic maps. Due to their pseudo-random behavior, chaotic systems can mimic randomness, which is essential for addressing complex problems through feature selection. In this methodology, the logistic equation is employed to generate periodic sequences of varying lengths within the TSO evolutionary algorithm. Experimental evaluations are conducted on 12 datasets from the UCI repository. On average, modified TSO improves classification accuracy by 15% over traditional methods across the datasets tested. Experimental results demonstrate that the proposed method has improved classification accuracy, diversity, and convergence, while removing redundancy.