Abstract

Breast cancer remains one of the most common cancers among women worldwide. Early detection and diagnosis are critical for effective treatment and survival. Traditional diagnostic methods rely heavily on expert interpretation, which can be time-consuming and prone to human error. With the advancement of machine learning and convolutional neural networks, it is now possible to develop intelligent systems capable of accurately classifying breast tissues as normal, benign or malignant based on medical data. This project aims to develop a machine learning-based and convolutional neural network-based classification system that can assist healthcare professionals in making more accurate diagnoses. For this project four Machine Learning Classification models and three Convolutional Neural Networks were trained and evaluated on datasets of diagnostic data and ultrasounds respectively. Machine Learning models showed very high performance across all metrics with XGBoost and Logistic Regression having the best performance with scores above 96% in all evaluation metrics. There was greater variability in performance across the CNN models. The best performing CNN model was the LeNet-5 model with an accuracy of 99.36%. The CNN models were used to make predictions on a set ultrasound images, all predictions were accurate.

Key words: Breast Cancer, Machine Learning, Convolutional Neural Network, Tumour, Classification, Dataset, LeNet-5, XGBoost, Logistic Regression