Linear regression with normal equation provides an analytical solution to approximate the theta value by using following equation:

Key points:

• There is no need of feature scaling
• There is no need of learning rate
• It can be computationally expensive if the features are very large matrix

References