Logistic_Regression

Repo having easy to medium Logistic Regression models

The equation of Linear Regression is y' = wx + b

In Logistic Regression the idea is to add a probabilities to the model

            1      
y` = ------------------
      ( 1 + e ^ (-wx+b) )

Error Calculation Method :

Cross Entropy

STEPS

TRAINING

• Initialize weights as zero
• Initialize bias as zero

GIVEN DATA POINT:

• Predict result y`
• Calculate error
• Use gradient descent to figure out new weight and bias values
• Repeat n times

TESTING:

• Put values from the data point into equation
• Choose the label based on probability