Supervised Learning-Based Prognostic Modelling For Early Detection of Breast Cancer

Abstract

Breast cancer continues to be among the leading cause of cancer-related mortality in women globally, and therefore there is the need to have more accurate diagnostic techniques for both diagnosis and treatment of the condition. Manual diagnoses that are currently being carried out using technologies like mammography, MRI, and ultrasound tend to be relatively slow, subjective and with some level of inter-observer variability. Therefore, this project proposes to come up with an advanced prognosis system that makes use of supervised machine learning techniques to ensure accuracy of breast cancer diagnosis as well as minimize the cost involved. Some of the machine learning algorithms like Random Forest and Decision Tree will be utilized along with computer aided diagnosis tools to determine various features that can include the tumor size, shape and texture as well as nuclear features of the cancer. Preprocessing steps included normalization, handling of missing values and feature selection before carrying out model training and testing. From the experimental results, it can be noted that the Random Forest model performed better than the decision tree model with higher accuracy and lower number of false positives and negatives.