Establishing Accurate Machine Learning Models to Support the Prediction of Prostate Cancer

Prostate Cancer (PC) is a austere disease amid men and universally it is the second most threat for the men life. Like other cancers, PC diagnosis also requires attention at earlier stage. For early prediction of PC, in this proposal, the dataset of 100 patient's magnetic resonance images (MRI) areconsidered and analysed using the experimental setup. The modality in image processing techniques provides accurate anatomical view of early stage of PC. Machine learning (ML) models can augment current clinical systems and raise patient care by pointing out areas of interest for doctors. ML techniques have enabled classification and segmentation of various types of PC's in a better way, providing objective results with enhanced precision. This work exploits familiar ML models for the classification and detection of early stage of PC from the patient dataset. Our implementation involves seven ML models; Logistic Regression, XGBoost, LightGBM, Random Forest, AdaBoost, Support Vector Machine (SVM) and KNN to evaluate the performance using different metrics and model is paired with a different classifier. Out of all these models, the Logistic Regression had out performed with a maximum of 90% accuracywith clear Receiver Operating Characteristic (ROC) curve and High AUC score.