----------- README -----------
To run the code type the following command:
$python run_models.py

You may need to install some libraries if you have not them already installed:
- numpy
- scipy
- sklearn
- nolearn

The code will go search for the audio directory for the .wav file and matching .txt files.
Matching is done based on names of file and locations.

It will output performance for DBN model. If you want to test another model,
go to run_models.py and look for the following lines to uncomment:
#compute_gmm(data_features, target, n_classes)
#compute_hmm(data_features, target, n_classes)

You can also easily chnage the number of features taken into account. For this look
in the run_models.py file for the following line and change the value for the number
of features you want to consider.
#nb_features = 13

For more tuning on selection of features & labels to give to the model, please refer to
the get_features.py file, where you will find boolean conditions that you can set on/off as
you wish to test the different possible combinations and their impact on the performance.