It is difficult to classify human emotions based on the electroencephalogram (EEG). In this project, the Support Vector Machine (SVM) is used to classify emotions trained on the DEAP dataset to predict emotions based on arousal-valence dimension. The DEAP dataset is a multimodal dataset for the analysis of human affective states [1]. The electroencephalogram (EEG) and peripheral physiological signals of 32 participants were recorded as each watched 40 one-minute-long excerpts of music videos. Each participant was asked for doing self-assessment to rate each video in terms of the levels of arousal, valence, like/dislike, dominance and familiarity. For 22 of the 32 participants, frontal face video was also recorded [1].

The data of each participant has 32 EEG channels with four labels (valence, arousal, dominance, liking) (see details on DEAP Dataset). These values is decomposed to five features by using Wavelet Transform [2] These five features are based on the frequency of the signals:

1. Delta (< 4 Hz)
2. Theta (4-7 Hz)
3. Alpha (8-15 Hz)
4. Beta (16-31 Hz)
5. Gamma (> 32 Hz)

Two ranges of frequencies are removed (0-0.5 Hz to avoid the artifacts, and near 50 Hz to reduce the effect of power line on signals).

The traditional SVM with default parameters as a classifier is used to be a model. However, other classifiers, such as neural networks and deep learning models should be applied in comparison in future work.

To install dependencies of this project:

python setup.py

To start the program:

1. Download DEAP Dataset (you need to sign disclosure). We will use data_preprocessed_python, make sure you download the right one.
2. Put the entire folder in the same working directory
3. Make sure you edit the path, DATASET_PATH variable in load_deap.py regarding the path of data
4. Run load_deap.py to get the data ready for training
5. Run svm_clas.py to train the model

1. Koelstra, S., Muhl, C., Soleymani, M., Lee, J.S., Yazdani, A., Ebrahimi, T., Pun, T., Nijholt, A. and Patras, I., 2011. Deap: A database for emotion analysis; using physiological signals. IEEE transactions on affective computing, 3(1), pp.18-31.

Disclaimer: This project originally published by Raghav714. I modified the code to be compatible with Python3, updated the libraries, and refactored the code and document accordingly. This project is just-for-fun.