Preprocessing dataset in IDS

This code transfer nominal features into numeric and then normalize the whole dataset using min-max

Most of classifiers in IDS area, especially artificial intelligence such as SOM, handle only numeric dataset and ignore the symbolic features. Therefore, in this section I present a simple version of matlab code that transfers nominal features in KDD dataset into numeric. In addition, after transferring, it normalizes the dataset using minimum maximum normalization to scale all features into [0,1] to avoid dominance and feature Impact.
The principle idea is considering all KDDcup features as random variables. Hence, this code use the probability mass function to transfer all nominal (symbolic) features, e.g. protocol_type, into numeric ones.

Unfortunately, you need to define the input excel file manually at the beginning, then you need to determine which column is nominal ...
The code is well commented and I think it will be easy to understand