Greylag goose optimization and multilayer perceptron for enhancing lung cancer classification

Area of Science:

• Computational biology
• Medical informatics
• Machine learning in healthcare

Background:

• Lung cancer poses a significant global health challenge, necessitating accurate diagnosis and staging for effective treatment and prognosis.
• Computational methods, particularly machine learning, are increasingly vital for enhancing lung cancer classification precision.
• Feature selection is a critical step in machine learning for handling complex datasets and improving model performance.

Purpose of the Study:

• To evaluate the efficacy of the Greylag Goose Optimization (GGO) algorithm in refining feature selection for lung cancer classification.
• To compare the performance of GGO with other binary optimization algorithms in enhancing classification accuracy.
• To assess the potential of GGO as a tool for improving lung cancer diagnosis.

Main Methods:

• Data preparation techniques including scaling, normalization, and gap factor handling were applied.
• The Greylag Goose Optimization (GGO) algorithm was employed for feature selection in lung cancer classification.
• A hybrid approach combining GGO with the Multilayer Perceptron (MLP) model was utilized.
• Statistical analysis using Wilcoxon signed-rank test and ANOVA was performed to validate results.

Main Results:

• The GGO algorithm demonstrated superior performance in feature selection compared to other binary optimization algorithms.
• The hybrid GGO + MLP model achieved a high classification accuracy of 98.4% for lung cancer.
• Statistical analysis and graphical illustrations confirmed the adequacy and efficiency of the proposed method.

Conclusions:

• The Greylag Goose Optimization (GGO) algorithm is effective in selecting optimal features for lung cancer classification.
• The GGO + MLP hybrid model shows significant promise for accurate and efficient lung cancer diagnosis.
• This computational approach offers a valuable advancement in the field of lung cancer detection and management.