Develop neural network application in C#.
We have would like to use NIR spectrometer to identify various types of material. The application should have a learning mode, where we can feed NIR spectrographs of the material. and then in output mode match the NIR and detect the material type.
Programming: C# .Net (Microsoft)
Development of Multiple class classification neural networks.
We have data same as given sample data for about 10 base materials and in different compositions.
The signature attribute is peak position and peak height.
The algorithm should contain following functions:
1. Function for peak position and peak height: the function should be able to obtain peak position (for both positive and negative peaks (i.e. local maxima and minima).
2. Function for Training models: once called this function should call the peak function obtain the peaks and peak positions for every dataset available create and train a multi classification neural network by the obtained data. And save the hidden node data at a memory location.
3. Function for prediction: The prediction function should take a single data entry as input. Obtain the peak positions and peak heights. Run data through the trained network. And identify in which class it belongs.