Linear SVM for Classifying Breast Cancer Data Encrypted Using Homomorphic Cryptosystem
Anny Kartika Sari, Faisal Malik Widya Prasetya · 2019 5th International Conference on Science and Technology (ICST) · 2019
Support Vector Machine (SVM) is widely used for data classification in many fields, including in biomedical area. While data privacy is a requirement for patients' health records, most existing classification models that are based on SVM cannot be operated on encrypted data. With the increasing use of cloud system, the need to process encrypted data cannot be avoided, especially for data stored in cloud system. This paper proposes a classification model that is based on Linear SVM. The model works on data that is encrypted using Paillier homomorphic encryption. Modification on Linear SVM classifier must be conducted to suit the requirement. Based on the evaluation, the accuracy, precision and recall of the proposed model is as good as those in the model that is applied on un-encrypted data. In terms of processing time, the application of modified Linear SVM classifier on encrypted data does not increase the time complexity.