Breast Cancer Prediction Using a Deep Learning Algorithm on the Cloud Medical Data

Recent strides in data generation have given rise to an unprecedented surge in information accumulation, marking the advent of the big data era. Unfortunately, conventional machine learning algorithms find it challenging to adeptly manage the distinctive characteristics presented by big data. This study is centered on the task of breast cancer prediction in the midst of this data-rich landscape. Specifically, we delve into two pivotal data types: gene expression (GE) and DNA methylation (DM).Our objective revolves around unlocking the potential of Deep Learning algorithms, both when applied individually to each dataset and when they come together in synergy. For this purpose, we've elected MATLAB as our primary platform. At the heart of breast cancer prediction, we employ the Convolutional Neural Network (CNN) algorithm, renowned for its proficiency in pattern recognition and feature extraction. Our study is underpinned by a comprehensive analysis that encompasses GE, DM, and the amalgamated GE and DM datasets.Furthermore, we gauge the performance of the CNN algorithm by placing it head-to-head with the Random Forest (RF) algorithm, a widely adopted choice in machine learning for classification tasks. The outcomes of our experiments vividly underscore the superiority of our approach, especially when we harness the potential of the GE dataset. This particular configuration yielded the highest accuracy and the most cost-effective results among all the classifiers considered. These findings serve as a beacon, illuminating the remarkable efficacy of Deep Learning techniques in the domain of breast cancer prediction, particularly when set against the backdrop of big data. This study underscores the remarkable potential of Deep Learning algorithms as they navigate the complex healthcare landscape and promises more accurate, cost-efficient diagnostic tools on the horizon.