Neural network backpropagation algorithm

Please I am going to desig a simple neural network with the following dimensions: 26 elements and 100 samples (26*100 input matrix ), 26 output neurons and only one sample (26*1 target matrix) ,1 hidden layer with 10 neurons .

problem description : i need a valid code for in-core fuel management optimization (nuclear engineering problem ) in which number of samples each consists of 26 element are entered as input matrix and only one of the samples is chosen as output (26*1) based on a reference sample .

 please I need a valid code to execute this design using the feed-forward back propagation algorithm;especially how to set my targets. Also I need parameters to check the network performance especially error rate.

Answer 1

[Greg Heath]

Regression and classification Neural Nets are based on matching I-dimensional input vectors with corresponding O-dimensional output vectors.

Are you able to reconfigure your problem into this scenario?

[Bachtiar Muhammad Lubis]

if case for extracting all text in document image o word docume(.doc) . are the inputs character in that document ( cropped image) or something else ?int