Algorithms of Neural Networks written in MATLAB
The code takes the input vectors, weights, learning constant and plots the updated weights at each stage
The code multiplies two matrices
This matlab code trains the weights for Bi-directional Associative Memory Network for english alphabects when made in a matrix of 5x3
Specific example of neural net based on competition. It can be used as a subnet to pick the node whose input is the largest.
This matlab codes takes n number of input neurons with the following parameters: -> radius of region of interconnections -> radius of region with positive interconnections -> constant c1 -> constant c2 -> external signal. The code performs the mexican hat algorithm on these input neurons for and desired number of times.
These nets can be used to find exemplar that is closest to the bipolar input vector x.
Kohonen Self-Organizing Maps also known as topology preserving maps algorithm has been demonstrated with this code.
The code shows how Linear Vector Quantization algorithm works. For now the code classifies 2 classes. Further improvement to the code will be made.
The code shows the implementation of perceptron learning algorithm for Logic gates. For the initial stage, AND Gate has been implemented whose input valus and target output can be easily modified in the code. It takes binary input and fetches bipolar output. This can also be modified.
The code shows the implementation of madaline learning algorithm for the logic gates. For the initial stage, the XOR Gate has been implemented. The target input can be easily modified in the code as per requirement of the gates. The code takes bipolar input and fetches bipolar output (which can be modified as per the need).