In logistic regression, decision boundary is a line that separates class A from class B
let's take a look at an example:
from the example plot above, the blue line is the decision boundary, separating y=1, from y= 0
To recap from yesterday, the logistic regression's function is g(z), and it's going to output a value between 0 to 1, it may output values such as 0.7 or 0.3
What if we want the learning algorithm to predict value of 0 or 1? We will set a threshold in this case, and let's say we want to set the threshold to 0.5:
to get a prediction (y=0 or y=1) from a logistic regression model, we can use the following heuristic: if f(x) >= 0.5, predict y=1, if f(x) < 0.5, predict y=0
If we were to plot this, it will look like this:
Comments