A Study on Early Prediction of Lung Cancer Using Machine Learning Techniques
V. Nisha Jenipher, S. Radhika
Abstract
Machine learning techniques are being used in cancer research for more than a decade. Nowadays, Machine Learning Algorithms (ML <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a</sub> ) can contribute significantly to the area of Lung cancer (LC) research. LC accounts for the highest mortality rate across the globe, hence early prediction and classification of cancer cells can increase the survival rate substantially. Though there are many algorithms used in the field of neurology, radiology, oncology for LC prediction, ML <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a</sub> outperforms those algorithms due to their accuracy and efficiency. This study first focuses on the workflow methodology used by ML <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a</sub> for early prediction and classification of LC. The methodologies include selecting the input data, preparing the data, feature selection and extraction, training and testing the data, and selecting the best ML technique. Second, a survey report of the ML algorithms used in LC and their methodologies is presented. Third, the performance metrics such as Accuracy, Sensitivity, Specificity, Precision, F1 Score, Root Mean Square Error (RMSE), Confusion Matrix, Area Under the curve (AUC) - Receiver Operating Characteristics (ROC) curve, Precision-Recall (PR) curve with different ML <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a</sub> are analyzed. Finally, this study also covers the parameters used in constructing an efficient and accurate ML model for the early prediction of LC.