ABSTRACT
Lung cancer is considered a cause of increased mortality rate due to delays in diagnostics. There is an urgent need to develop an effective lung cancer prediction model that will help in the early diagnosis of cancer and save patients from unnecessary treatments. The objective of the current paper is to meet the extensiveness measure by using collaborative feature selection and feature extraction methods to enhance the dendritic neural model (DNM) in comparison to traditional machine learning (ML) models with minimum features and boost the accuracy, precision, and sensitivity of lung cancer prediction. Comprehensive experiments on a dataset comprising 1000 lung cancer patients and 23 features obtained from Kaggle. Crucial features are identified, and the proposed method's effectiveness is evaluated using metrics such as accuracy, precision, F1 score, sensitivity, specificity, and confusion matrix against other ML models. Feature extraction techniques including Principal Component Analysis (PCA), Kernel PCA (K-PCA), and Uniform Manifold Approximation and Projection (UMAP) are employed to optimize model performance. PCA evaluated the DNM accuracy at 96.50%, precision at 96.64% and 97.45% sensitivity. K-PCA explained the DNM accuracy of 98.50%, precision rate of 99.42%, and 98.84% sensitivity and UMAP elaborated the DNM accuracy of 98%, precision of 98.82%, and 98.82% sensitivity. The K-PCA approach showed outstanding performance in enhancing the DNM model. Highlighting the DNM's accurate prediction of lung cancer. These results emphasize the potential of the DNM model to contribute positively to healthcare research by providing better predictive outcomes.
