Dual-Model Machine Learning Validation Dataset for Early Breast Cancer Triage in Rural Bihar, India

This dataset provides a validated corpus for a dual-model machine learning framework designed to revolutionize early breast cancer screening in the resource-constrained rural landscapes of Bihar, India. It integrates high-fidelity image detection scores from a MobileNetV2 CNN fine-tuned for edge computing with clinical risk predictions from a Random Forest Classifier that evaluates epidemiological factors like parity, family history, and BMI. The data validates a specialized triage logic optimized for a 91.2% sensitivity rate, ensuring that community-embedded ASHA workers can provide immediate, culturally sensitive, and life-saving referral decisions without relying on continuous internet connectivity or centralized hospital infrastructure. By bridging the gap between advanced CAD diagnostics and localized field application, this dataset serves as a technical foundation for reducing late-stage diagnosis rates and destigmatizing breast health in conservative rural heartlands.