The rbf and other two neural network algorithms that I just learned recently, rbf is a radial basis function, rbf has strong approximation ability, classification ability and learning ability, and can approach nonlinear functions infinitely; rbf neural network consists of a Input layer, hidden layer, output layer, simple structure and fast convergence speed;
The rbf (radial basis function) neural network, the activation function of the rbf neural network takes the Euclidean distance as the independent variable, and the function is generally a Gaussian function; it can be 
seen from the picture that when the dist is smaller, the R function is larger, and dist is the weight The Euclidean distance between the value vector w and the input vector x; b is the threshold to adjust the sensitivity of neurons; the hidden layer of rbf is a nonlinear mapping, which converts data from low-dimensional to high-dimensional, and then easily linearly separable from high-dimensional ;The output layer of rbf is linear, and it is adjusted linearly; the input layer of rbf is only to transmit data; the activation function is in the hidden layer of rbf, and the output layer is a linear combination of data;