1. The activation function of the BP network must be differentiable everywhere. 2. The area divided by the S-type activation function is a nonlinear hyperplane area, which is a relatively soft and smooth arbitrary interface. Therefore, its classification is more accurate and reasonable than linear division, and the fault tolerance of this network is better. Another important feature is that since the activation function is continuously differentiable, it can be strictly calculated using the gradient method. 3. In general, the BP network structure uses the S-type activation function in the hidden layer and the linear activation function in the output layer.
You Might Like
Recommended ContentMore
Open source project More
Popular Components
Searched by Users
Just Take a LookMore
Trending Downloads
Trending ArticlesMore