Transparent Lung Cancer Risk Assessment via Hybrid Machine Learning Algorithm Leveraging Interpretability

The increasing prevalence of cancer in today's world makes it increasingly difficult to save lives. The basic reason behind occurrence of lung cancer is smoking and hormonal factors. The cause of lung cancer depend upon many factors but biopsy helps to detect lung cancer at early stage. The main cause of affect the Therefore, early detection of the disease is crucial for providing effective treatment to patients. Lung cancer, in particular, has a high mortality rate that affects both men and women. In this type of cancer, cells in the lungs grow uncontrollably. The percentage of lung cancer disease has been found in Males more than female. The ratio may be 65% in Male and approximate 40% in Females. Early detection of lung cancer can significantly impact a patient's chances of survival. The aim of this paper is to predict the disease by examining various parameters that contribute to its early onset, utilizing different machine learning techniques such as Linear Regression and Random Forest and hybrid approach. Hybrid approach performed well to lung cancer prediction. The redactions have been depending upon by combines the best features of all applied algorithms. These calculates best features improve a robust testing and emphasis interpretability. Also work well proposed hybrid model on unseen data. The evaluation using a confusion matrix revealed accuracy more than 97%. The time taken to train model is less than 0.4seconds.