No self-introduction
Loading
Profile
Follow
codes (2)
Autoregressive model
no vote
Time varying autoregressive model (TVAR) matlab code
MrDou 1 0
1
Error analysis of MATLAB
no vote
The determination of the number of hidden layer nodes is a very important link in the design of neural network. A two-layer BP network with infinite hidden layer nodes can realize any nonlinear mapping from input to output. But for the limited input to output mapping, it does not need infinite hidden layer nodes, which involves how to select the number of hidden layer nodes. However, due to the complexity of this problem, a good analytical formula has not been found so far. The number of hidden layer nodes is often determined according to the previous design experience and their own experiments. Generally speaking, the number of hidden layer nodes is directly related to the requirements of solving the problem and the number of input and output units. Moreover, if the number of hidden layer nodes is too small, it can not produce enough connection weight combinations to satisfy the learning of several samples; if the number of hidden layer nodes is too large, the generalization ability of the network becomes poor. There are several methods to determine the number of nodes in the hidden layer: & nbsp; 1) if the sample function required to be approximated changes violently and fluctuates greatly, it is required to have more adjustable connection weights, so that the number of nodes in the hidden layer should also be more; & nbsp; 2) if the prescribed approximation accuracy is high, the number of hidden layer elements should also be more; & nbsp; 3) We can consider adding fewer hidden layer units at the beginning and gradually increasing them according to the later learning situation, or adding enough hidden layer nodes at the beginning and deleting the less effective connection weight and hidden layer nodes through learning
MrDou 0 0
1
No more~
Follows0
Fans0