.

in principle, with a standard supervised feedforward ann, you do not _need_ to normalize the inputs : the training process will adapt the input weights to the scale of your data ; it will however take time and you can hope to speed-up the training by normalizing your inputs in a reasonable manner (that is a manner which takes the distribution into account : for instance, blind min-max normalization may hurt your training badly if you have a very large outlier)

the only thing you _must_ to normalize is your target data so that it fits in the range of your output layer function !

.

You don't have a problem doing that. The first 20 variables are using a common measure in the range [0,1]. Put the last variable in the same range by using any membership function of your choice, even an inverse function can do that.

In addition to what is said by David Nwanze and Fabrice Clerot, you can choose one of the following techniques in normalizing

1. Z transformation

2. Range Based Transformation

But use all the attributes while normalizing it improves data set consistency.

The 20 variables are normal, only the 21st should be normalized

you can normalize (standarize) the last variable by using a normalization equation

Your Variable - Minimum Number in your set / Maximum - Minimum Number

Hamzah Al Najada answered you. If your variable = min you get 0. If your variable=Max you have 1. All other values will be between 0 and 1.

Thank you all for your answers.

I found another track, for which I would not need to normalize the data

at the same time i  go back to the answer of Dr  David Nwanze , can you recommend any reading on the use of membership function for normalization 

Thank you

Any book on fuzzy logic will do. Normalization is the basis for fuzzy computations and allows you to combine data in the same scale. Everything depends on your knowledge engineering skill to ferret out information from experts on the way they solve their problem. You work with experts in the area of your application to determine the threshold of your data.

I would recommend this video: 16 6 Machine Learning Implementation Detail Mean Normalization)  (Link attached)

Prof. Andrew Ng

Stanford University

http://www.youtube.com/watch?v=TqeIwYxV398

Hi Nihad,

Try this simple routine for normalization:

% To normalise data between -1 and +1

for a=1:m,

dis(a,:)=(2./(maxs-mins) .* (dis(a,:)-mins))-1;

where 'dis' is your parameter matrix with 'm' lines. Maxs and Mins are the max and min values of your array.

I hope this helps.

Luiza