Gaussian Mixture Model based Hybrid Machine Learning for Lung Cancer Classification using Symptoms

Abstract

Being a fatal disorder, lung cancer becoming a primary reason for mortality in people who are affected with various symptoms. This implies that there is always a necessity in the medical field to have a promising approach for detection and timely treatment for such disorders. Also, it is required to be done at an earlier stage to attain a reduced mortality rate among cancer patients. The work intended to propose a hybrid machine learning (ML) strategy for the classification of lung cancer. The approach incorporates both Non-Linear Regression (NLR) and Gaussian Mixture Model (GMM), combinely termed as NLR-GMM algorithm. The algorithm takes the key advantages of both machine learning models for better classification of lung cancer data. For this, the work employs the lung cancer dataset constituted using its symptoms. The data set is preprocessed and visualized for analysis. Then classification is performed using the proposed hybrid ML approach which provides a maximum performance of 92.88% of classification accuracy. The results are compared with the existing ML algorithms such as Gaussian Naïve Bayes and K-Nearest Neighbor algorithms for checking the proposed strategy.