Five minutes per day machine learning: use gradient descent to complete the fitting of multi feature linear regression model

This paper mainly ：

Complete gradient descent of multiple features , Similar to univariate , It's updated θj More

Multi characteristic hypothetical function

ℎθ() = θ0 + θ1x1 + θ2x2+…+θx

Multi characteristic loss function

With the loss function, we naturally want to find the minimum value of the loss function , Then the way to use it is gradient descent

We can get the formula of gradient descent by expanding the partial derivative as ：

among x0(i)=1, We start to initialize and select a series of parameter values θ, Then update at the same time θ0、 θ1、θ2……θn, Achieve gradient descent , To find the final θ, Find the hypothetical function that best fits the data , This is the calculation process of multivariable gradient descent , This is the same as univariate .