Emotional behavior analysis based on EEG signal processing using Machine Learning: A case study

Based on a well-known benchmark, a comparison between the present study and the literature was carried out. This paper investigates a variety of Machine Learning (ML) and Deep Learning (DL) algorithms for classifying emotional events using EEG brainwave data. The contribution of this paper occurs in the data processing phase more precisely at the classification level to predict human emotions either positive, neutral, or negative from EEG signals after applying several algorithms and techniques. According to Bird’s findings, RF augmenting with InfoGain information outperforms Adaptative Boosted LSTM, Adaboosted MLP, and nonboosted DEvo MLP. During the classification phase, we used different classifiers such Random Forest (RF), XgBOOST, NaiveBayes (NB), Decision Tree (DT), Linear RegressionCV (LRCV), Support Vector Machine (SVM), Linear Regression (LR), and Convolutional Neural Networks (CNN) to improve classification performance. They attained an overall accuracy of around 96,88%, 96,41%, 95,47%, 94,06%, 90,00%, 89,06%, 88,91%, and 52,66% respectively. As a result, we find that InfoGain consistently improves RF’s performance in dealing with data and outperforms other classifiers. On the other hand, the inefficiency of CNN can be explained by the lack of a big amount of data.