In the tutorials Logistic Regression model is built using the Tenseal for a Classification problem. I'm planning to implement with Linear Regression using Tenseal for a Regression Dataset. I'm achieving a good accuracy score for the first 2 parts i.e. 1. Linear Regression model built with Plain data and tested on the plain data, 2. Encrypted Evaluation. But for the 3rd main part where we are building the model on top of Encrypted Data getting various issues. How Can I replace the existing sigmoid function in the encrypted model to something else so that the model works fines on top of the Encrypted data and gives good results?
HibaHajar
commented
11 months ago
Hello,
In the tutorials Logistic Regression model is built using the Tenseal for a Classification problem. I'm planning to implement with Linear Regression using Tenseal for a Regression Dataset. I'm achieving a good accuracy score for the first 2 parts i.e. 1. Linear Regression model built with Plain data and tested on the plain data, 2. Encrypted Evaluation. But for the 3rd main part where we are building the model on top of Encrypted Data getting various issues. How Can I replace the existing sigmoid function in the encrypted model to something else so that the model works fines on top of the Encrypted data and gives good results?
Thanks in Advance!!!