Privacy-preserving analytics : secure logistic regression
Much data and information have been collected about us from all aspects of our life. Sometimes, we need to do analysis on this data without violating the privacy of individuals. In this project, we present a cryptographic library that can be used to do logistic regression under encrypted data. The e...
Saved in:
| 主要作者: | |
|---|---|
| 其他作者: | |
| 格式: | Final Year Project |
| 語言: | English |
| 出版: |
2019
|
| 主題: | |
| 在線閱讀: | http://hdl.handle.net/10356/77126 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 機構: | Nanyang Technological University |
| 語言: | English |
| 總結: | Much data and information have been collected about us from all aspects of our life. Sometimes, we need to do analysis on this data without violating the privacy of individuals. In this project, we present a cryptographic library that can be used to do logistic regression under encrypted data. The encryption scheme used is a multiparty computation based on Exponential ElGamal. A special type of multiplication gate, the conditional gate, helps in the realization of the library. An implementation of the library usage on predicting the severity of heart disease based on the encrypted patient’s attributes is also presented along this project. |
|---|