Understand logistic regression and logit function in depth.
Updated: Nov 2, 2022
Logistic Regression is a classification algorithm where we only have two choices either true or false.
A categorical target variable is used for doing prediction, and a categorical variable is one that represents characteristics and can't be measured or counted.
Here the probability lies in a range of 0 and 1.
If the probability lies between 0 to 1, then why can't we use Logistic Regression?
Because it is susceptible to outliers, due to outliers threshold points can shift.
Due to the interpretation of the model.
Since it is a type of classification problem, but still used Regression in its name, which is because we need continuous values in the range of 0 to 1.
We need to modify the linear regression equation for achieving this task and need to implement a new logit function.
It is a function that has a range between 0 and 1, irrespective of any input it will always give output between 0 and 1.
If (mx+c) is high then the overall Y value comes out to be 1.
If (mx+c ) is low(or negative) then the overall Y value comes out to be 0.
Hence we are getting a new S-shaped curve.
Example: If the probability comes out to be 0.8 then it belongs to class 1.
Therefore if the probability is greater than 0.5 then it belongs to class 1 else class 0.