Greylag Goose Optimization (GGO) Algorithm for Classifying Lung Cancer Disease

Abstract

It is to this effect that this paper proposes a Greylag Goose Optimization-based algorithm for the improvement of accuracy in case classification for lung cancer. The input data used for this study is prepared by scaling, normalization, and removal of null values. To get an optimal subset of features to improve the classification accuracy, the binary version of the GGO algorithm is compared with six other optimization algorithms: bSC, bMVO, bPSO, bWOA, bGWO, and bFOA, proving the efficacy of bGGO in feature selection. Multi-classification using many classifiers predicts MLP as the superior one with an accuracy of 91.8%. Hyperparameter tuning using GGO enhances the accuracy of MLP to 98.4%. Statistical evaluation with ANOVA and Wilcoxon's signed-rank test establishes the outcome to be highly significant (p < 0.005). The hybrid method of GGO + MLP reveals better robustness and efficien.