Module to handle input/output operations and to normalize inputs. Data set is read to respective variables (see loading the data set in Section 2). The data set is organized in two files, one is for holding the inputs to the network and the other for the corresponding outputs of the network.
inputfilename variable content is used to open
the file containing the network inputs, and read the contents as a matrix of
size numInputs
nOfSamples. The data is read into the variable
inputs. Note that the file has the data in a transposed manner, i.e.
columns of length numInputs and nOfSamples
rows.
outputfilename variable content is used to open
the file containing the desired network outputs, and read the contents as a
matrix of size numOuts nOfSamples .
The data is read into the variable desiredOutsL2 .9
After the inputs are read, according to the variable normalizeInputs , the input values are normalized. This is done by first finding the largest input value that exist in the data set and then dividing all inputs by that number, hence the maximum input value will be equal to 1.