Skip to content

Batch mode

Jump to bottom
Simone Maurizio La Cava edited this page May 1, 2020 · 3 revisions

The batch mode allows you to execute all your studies without having to execute manually each step for all the measures.

You will be only required to insert a text file which contains all the wished parameters.

So, let's have a look of this file.





The file

The .txt file has to contain all the parameters use for the study.

Here, you can choose the steps to do, by setting the relative parameters to true or false, and set all their parameters (for the steps which you want to jump, you can leave empty the related parameters or fill them with any value).

Furthermore,you will be able to make the study, using the same parameters, for all the wished measures, to make some studies by using more spatial subdivision, and insert more parameters which you cannot use in the guided mode for the classification:

  • RF_Pruning, which limits the maximum depth of the Random Forest classifier (it may prevent the overfitting) and can take yes or no values
  • RF_MinimumClassExamples and NN_MinimumClassExamples, which allow you to set the minimum number of samples required in the randomly selected training dataset in both classifiers (1 by default)
  • RF_Rejection and NN_Rejection, which allow to set the minimum threshold value of probability of the most likely class (if the probability which a sample belongs to a class is less than this value for each class, the sample is simply rejected), and take a value between 0 and 1

The batch mode allows also to set the number of bins to use in the histogram analysis (high, medium and low values are permitted).

However, the order of the parameters is not important: the file reader of Athena will only search for the parameter name in order to set its value.

Note that you can generally set a number or a string as value of a parameter, but there are some parameters in which you have to insert more values, such as cf (which represents the cut frequencies), totBand (which represents the list of cut frequencies of the total frequency bad, it is only used in the PSDr extraction), the measures field (which represents the list of the measures to extract and study) and the fields relative to the spatial parameters: in these cases you have to leave an empty space (with the space bar on your keyboard) between two inserted elements.

If you want to leave a parameter as its default value, you have just to insert none as its value.


In the main directory you can find an example_file, which explain the meaning of the various parameters and the step in which they are used.

Consider also that when you compute a study on some measure, it will be created a text file which contains the parameters used in its extraction and the subjects' file and the eventual locations' file used, and you can also use this file (maybe modified before in order to compute also the following steps) in the batch mode.





The interface

Now we can go to the batch mode's interface: here we have just to insert the name of the text file, to use as parameters file and the Athena will automatically execute it, from the extraction of the measures to the analysis and the classifications.

Here you can also decide to try all the available classifiers: in this case the resulting file containing parameters and performance will be renamed in order to allow identify the classifier from which it come from.

Remember that you can always open this page of the wiki by clicking on the logo button, in order to open it in a web browser.

Now you can RUN studies and experiments, change the parameters by only changing the text file, launch the study and leave it work alone while you are making other stuff.

If you want instead to return to the guided mode, or analyze variations during the time of the signals and their sprectum, you can click on the < Back button in order to return to the initial interface and select another study mode.

Clone this wiki locally