In the Regression problem, it is assumed that the relationship between independent variables and the dependent variable is linear. But it may be possible that the relationship is non-linear.
When the relationship is non-linear, creating a straight line may not be the best decision boundary, we need to have a decision boundary that can fit the data well. Hence higher polynomial degree features are included in the model to make a non-linear decision boundary.
Polynomial Regression is also used to solve the problem of Underfitting.
Polynomial regression is a special kind of Linear Regression that includes the feature of n polynomial degree.
- y = predicted output
- b0 = intercept, bias
- x1 = independent variable
- = second polynomial degree of feature 1
- = third polynomial degree of feature 1
- = n polynomial degree of feature 1
- b1,b2,..,bn = coefficient of corresponding features
- n polynomial degree is the highest degree feature present.