Breast Cancer Prediction Using Machine Learning

According to the World Health Organization, Breast cancer is one of the most prevalent malignancies in women and continues to be a major cause of death worldwide, taking almost 670,000 lives in 2022 alone.Designing and implementing a web-based cancer prediction system that combines machine learning and deep learning models to offer precise and accessible diagnostic support is the main goal of this project.Within the system, two prediction models were created and implement.Thirty clinical characteristics (such as radius, texture, smoothness, and concavity) were used to train a Random Forest Classifier on the Breast Cancer Wisconsin Diagnostic dataset in order to categorize tumor as benign or malignant.Concurrently, Histopathological images were used to train a Convolutional Neural Network (CNN) for automatic image-based classification.A Flask web application with role-based dashboards, authentication, and user registration included both models.User information and prediction history are safely stored in SQLite database.By uploading medical photos or entering clinical data, end users can get predictions.The results are shown with confidence scores.The CNN performed well on image-based classification tasks, whereas the Random Forest model attained good accuracy for structured data.Both tabular and picture inputs can be handled by the hybrid, adaptable solution offered by the integrated system.This application demonstrates how computer vision can be used as decision support tool, potentially lowering diagnostic delays and assisting physicians in early cancer diagnosis, even if it is not meant to replace medical experts.