Hybrid FrFT-HS Framework for High-Accuracy Lung Cancer Classification from Gene Expression Data

Lung cancer classification from microarray gene expression data presents a significant challenge due to high dimensionality and redundant features. This study employs the Fast Fractional Fourier Transform (FFrFT) for dimensionality reduction, followed by Harmonic Search (HS) for optimal feature selection. Various classifiers, including Bayesian Linear Discriminant Classifier (BLDC), Gaussian Mixture Model (GMM), Naive Bayesian Classifier (NBC), Random Forest (RF), Decision Tree (DT), and Support Vector Machine (SVM) with Radial Basis Function (RBF) kernel, were evaluated for classification performance. Experimental results show that SVM (RBF) attained an impressive accuracy of 91.16%. This outstanding performance underscores the significant impact of proposed dimensionality reduction and feature selection pipeline in enhancing classification performance.