Comparative Analysis of Deep Learning Models for Breast Cancer Classification

Breast cancer is still one of the major causes of death for women, which emphasizes how critical it is to detect it accurately and early to improve outcomes. In order to create a trustworthy system for identifying breast cancer as either benign or malignant, this project makes use of deep learning techniques. Prior to improve the performance, optimization techniques like hyperparameter tuning and feature standardization are investigated in conjunction with sophisticated neural network topologies, such as Convolutional Neural Networks (CNNs) for image data and Feedforward Neural Networks (FFNNs) for structured data. To improve diagnostic efficacy, performance indicators including recall, accuracy, and precision are used. The class imbalances, dataset restrictions (such as the Wisconsin Breast Cancer Dataset), and uneven performance across several datasets are some of the issues the study tackles. The models are effectively trained and validated using Google Collab with GPU support, and a comparison of CNNs and FFNNs shows their advantages. The objective of this research work is to create a scalable and reliable diagnostic system that will aid in the diagnosis of breast cancer and enhance clinical judgement.