Machine-Learning Based Prediction of Lung Cancer

Abstract

The incidence of lung cancer has now exceeded all other types of cancer globally, making it the leading cause of cancer-related deaths. Compared to other types of cancer, lung cancer has a poor prognosis and high mortality rate, making it challenging for humans to accurately predict its incidence rates. The goal of the study is to implement machine learning techniques for the early detection of lung cancer, which can increase patient survival rates. To find the most important factors and predict the possibility of lung cancer, a variety of data mining approaches, including logistic regression, k-means, and apriori algorithms are used in this study. Age, gender, smoking habit, and medical history are a few of the factors included in the dataset used for the study. The logistic regression classifier shows an accuracy of 95% for the classification of lung cancer in patients. The simulation results obtained from the k-means clustering algorithm shows that the main causes for the possible occurrence of lung cancer are chronic diseases, fatigue, allergy, wheezing, alcohol consumption habit, and breath problem. Similarly, according to the association rule's findings, there is no chance of lung cancer developing in a non-drinker who is free of peer pressure, allergies, and wheezing issues.