It is difficult to look at the EEG signal and identify the state of Human mind. In this assign- ment, the SVM classifier is trained with Deap dataset to predict the state of mind. the state of mind is predicted in terms of valence, arousal. which can further be used to predict the state of mind in terms of expression.
In this assignment, the preprocessed data is used for training the classifier. Steps involve in training the dataset:-
The DEAP dataset consists of two parts:
In this assignment, Wavelet transform is used to decompose the each channel data into the five feature i.e • Delta (< 4 Hz) • Theta (4-7 Hz) • Alpha (8-15 Hz) • Beta (16-31 Hz) • Gamma (> 32 Hz) In this assignment, obtained the 7 decomposed values but we negalted the frequency whose range is in 0-0.5 Hz so that the artifcats are removed. The frequency whose range is near 50 Hz are removed to reduce the effect of power line on signals. finally, EEG band are obtained for each channel.
The dimension can be reduced using one of the below mention method:-
In this assignment, the classifier used is Support vector machine (SVM). we can also use other classifier or neural network to predict the values but the training efficiency is found to be nearly 98 percentage with SVM.