An Optimized Multistage Model for Lung Cancer Prediction with Machine Learning

Abstract

: Lung cancer accounts for a significant share of cancer-related deaths globally. It is one of the most prevalent and fatal illnesses. This paper presents an enhanced machine learning framework for diagnosing lung cancer. To to eliminate redundant and irrelevant features, a feature selection method inspired by natural phenomena called firefly algorithm is employed. Lung cancer patients are subsequently categorized using the K-Nearest Neighbors algorithm based on specific characteristics. The goal of the proposed model is to maximize computational efficiency while maintaining high diagnostic accuracy. Calculation metrics, such as accuracy, the number of features chosen, and computing efficiency, will be used to evaluate the model. This study supports improved treatment planning, better patient outcomes, and advances in early detection. Additionally, combining feature selection with classification reduces overfitting and improves prediction accuracy. Incorporating medical knowledge with computational intelligence provides a sustainable approach to developing scalable and clinically useful diagnostic systems. Experimental results show that the proposed model obtained 96.73% best accuracy. If applied in medical diagnosis systems, the suggested model can significantly increase the accuracy of lung cancer detection leading to operational treatment and an increase in patient survival rates.