Statistical Comparison of Valence-Arousal Classifiers from EEG on DEAP and MANHOB Datasets

This repository contains pretrained models, code and a report about a statistical study on different valence-arousal classifiers from EEG (Electroencephalography) signals on DEAP and MAHNOB datasets.

The report compares two classifiers (a deep neural network and a convolutional neural network) against DEAP and MAHNOB, and evaluates if there is a significant difference in performance between the two methods using some statistical tests (like McNemar's test and 5x2cv test).

Project structure

Project structure is explained in Section 7 of the report.