Linear Regression, Simple Linear Regression, Multiple Linear Regression, MSE, Cost function, Loss function, Objective function, Optimizer, Gradient Descent

• Regression

A prediction comes out within a range of any continuous value such as weight, age, speed.

 

• Linear Regression

It is used for finding linear relationship between dependent and one or more predictors.

x------>y  affect x : affect to y, predictor or independent variable y : subordinated by x, response or dependent variable

 

1. Simple Linear Regression

Presume one x.

y=Wx+b W : Weight, gradient b : Bias, intercept

 

2. Multiple Linear Regression

Presume more than one x.

y=W1x1+W2x2+...Wnxn+b

 

 

3. Cost function(Loss function, Objective function)

Error=real value-predict value

 

Minimize the error. W,b→minimize cost(W,b) 0 is the best!

 

4. MSE

Mean Squared Error.

 

5. Optimizer(Gradient Descent)

The process to find W,b for minimizing cost function.