Blockchain Enabled Federated Learning for Privacy Preserving Lung Cancer Risk Prediction

A new architecture for prediction of lung cancer risk using survey derived data by incorporating federated learning in combination with blockchain technology has been discussed. The proposed approach allows multiple healthcare institutions to collaboratively train machine learning models without the need to share sensitive patient data, which is a significant advancement in privacy and regulatory compliance. The dataset was segmented across three hypothetical hospitals, consisting of 309 records and 16 attributes. A local Random Forest classifier was created at each university, and models with accuracy above 85% were documented in a simulated blockchain utilizing the SHA-256 hashing algorithm. This ensures an immutable, verifiable provenance of model contributions. This architecture illustrates decentralization learning with high accuracy while maintaining user data privacy and trust in federated nodes. The results indicated a great opportunity for practical applications, especially in public health systems and joint clinical studies.