Hello everyone
In the logistic regression model we are told that the decision boundary in case of threshold 0.5 is w.x + b. Now here the w.x+b line is totally predicted by the linear regression and then we pass the computed result in the logistic regression function . Now adding points to the extreme far will cause the linear regression line to move which will in turn shift the decision boundary. So isn’t it the same problem we were facing at the first place because of which we shifted to logistic regression.
Not quite true.
The boundary is where the sigmoid of (w*x +b) = 0.5.