Interpretation via Feature Selection

This page is a supplemental material for the research paper titled:

Interpretation of Depression Detection Models via Feature Selection Methods

Submitted to IEEE Transaction of Affective Computing

Individual Feature Selection Methods

classification results in term of balanced accuracy (Acc.) and number of selected features (SF)

An interface is currently underdevelopment and will be made public soon.

All feature selection methods used in the paper can be selected.
Customized methods can be added by the user (following simple formatting).
Edited preferences for classification and selection process.
The results of the framework can be then used for interpretation and classification

Github link will be provided here when the GUI project is ready.

Page updated

Google Sites

Report abuse