LUNG CANCER PREDICTION USING GLCM WITH HYPERPARAMETER OPTIMIZATION APPROACH

Abstract

Abstract: Cancer is one of the most lethal illness professionals. This is responsible for an uncountable number of deaths throughout the globe.The concept of diagnosing lung cancer at an earlier stage piqued the interest of medical professionals.This research presents a novel method for the detection of lung cancer that is based on the processing of pictures obtained from CT scans.With the use of cases from the Lung Imaging Database Consortium (LIDC) database, we were able to assess the practicability of applying algorithms for the detection of lung cancer in this research. The primary purpose of this study is to determine if the tumours that are discovered in the lung are malignant or benign. This has been achieved using the GLCM feature extractor and the SVM classifier. The proposed research includes the hyperparameter optimization approach that has been proposed for use in identifying lung cancer in its early stages.  Results obtained with hyperparameter optimization has showed the overall accuracy prediction of 92.06% and more than 90% of accuracy in deciding tumour as malignant, benign, or normal. And area under the curve (AUC) value 0.993,0.996 and0.997 has been obtained for the classesmalignant, benign, or normalrespectively during the classification cancer.