<![CDATA[Breast cancer remains one of the leading causes of cancer-related deaths worldwide, highlighting the need for accurate and efficient diagnostic tools. This study focuses on implementing machine learning models, particularly Artificial Neural Networks (ANN), to classify breast cancer types based on genomic data. Using the METABRIC RNA Mutation dataset, the system combines a cloud-based backend with Supabase and an intuitive frontend built with Streamlit. To ensure data compatibility with the models, preprocessing steps such as standardization, label encoding, and one-hot encoding are applied. TensorFlow is used to load models saved in .h5 format, with two approaches tested: a 30-feature model achieving 99% accuracy and an average prediction time of 80 milliseconds, and a 6-feature model achieving 100% accuracy with a faster prediction time of 42.25 milliseconds. Prediction results are stored securely in Supabase, complete with timestamps for tracking and exported as PDF reports for easy documentation. Data security is prioritized through the use of API keys, JWT tokens, and Streamlit secret management to safeguard sensitive information. The integration of Supabase for backend processing, Streamlit for real-time visualization, and GitHub for CI/CD automation results in a scalable, reliable, and efficient system. This study presents a robust solution for breast cancer classification, providing real-time predictions, secure data handling, and a user-friendly interface suitable for clinical and research applications. Keywords— breast cancer classification, artificial neural network, genomic data, Supabase, Streamlit, real-time prediction, data security.]]>
Copyrights © 2025
