Analysis , investigation and detection of prostate cancer using machine learning

Abstract

Prostate cancer remains one of the most prevalent malignancies among men, necessitating accurate and early diagnostic systems. This research presents a comprehensive deep learning framework incorpo- rating Convolutional Neural Networks (CNNs), AdaBoost ensemble learning, and a Softmax classifier for effective prostate cancer detec- tion. The proposed methodology utilizes preprocessed image datasets to extract spatial features through CNN layers, followed by pooling operations for dimensionality reduction. These features are further optimized using AdaBoost to enhance classifier robustness. Finally, a Softmax layer performs binary classification between benign and malignant samples. Experimental results demonstrate significant im- provements in prediction accuracy, sensitivity, and specificity com- pared to traditional machine learning models. The proposed frame- work offers a scalable, interpretable, and efficient approach for clinical decision support in prostate cancer diagnosis.