Privacy-preserving Breast Cancer Prediction via Inner-Product Functional Encryption

Machine learning is increasingly being used in the medical domain in recent years. However, due to increasing privacy concerns and requirements imposed by privacy-related regulations, the deployment of machine learning solutions in the medical field poses a significant challenge. As a new type of advanced encryption primitive, functional encryption allows fine-grained access control and selective computation on ciphertext, making it very suitable for privacy protection machine learning scenarios. In this paper, we propose a privacy-preserving breast cancer predication system based on a provable secure inner-product functional encryption scheme. The medical server trains the logistic regression model based on breast cancer data, and then a client encrypts his test data using the inner-product functional encryption scheme, and sends the encrypted test data to the medical server for predication, and the medical server complete the prediction by running the decryption algorithm of the inner-product functional encryption scheme, and sends the predication result to the client. Experimental results show that the proposed system has a significant effect on breast cancer data privacy protection, and achieves a high accuracy, which are close to the accuracy of the non-privacy protection logistic regression model on breast cancer data.