Is Neural Network a combination of multiple linear functions?

I want to know is Neural Network a linear model or non linear?

What I think about NN is that It uses linear function to calculate the parameters and used non linear activation function to get the values between certain range.

At every layer of neural network, we are using two functions in sequence: first a linear function followed by the non-linear activation function. In mathematics, that is called “composition” of two functions. Since the second function is non-linear, that means the composition of the two functions is non-linear. Then when we have multiple layers, we are composing those functions: the composition of non-linear functions is also non-linear.

1 Like