We utilize a type of encryption scheme known as a Fully Homomorphic Encryption (FHE) scheme which allows for computation over encrypted data. Our encryption scheme is more efficient than other publicly available FHE schemes, making it more feasible. We conduct simulations based on common scenarios in which this ability is useful. In the first simulation we conduct time series analysis via Recursive Least Squares on both encrypted and unencrypted data and compare the results. In simulation one, it is shown that the error from computing over plaintext data is the same as the error for computing over encrypted data. In the second simulation, we compute two known diagnostic functions over publicly available data in order to calculate computational benchmarks. In simulation two, we see that computation over encrypted data using our method incurs relatively lower costs as compared to a majority of other publicly available methods. By successfully computing over encrypted data we have shown that our FHE scheme permits the use of machine learning algorithms that utilize polynomial kernel functions.
CITATION STYLE
Gribov, A., Horan, K., Gryak, J., Najarian, K., Shpilrain, V., Soroushmehr, R., & Kahrobaei, D. (2019). Medical diagnostics based on encrypted medical data. In Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST (Vol. 289, pp. 98–111). Springer Verlag. https://doi.org/10.1007/978-3-030-24202-2_8
Mendeley helps you to discover research relevant for your work.