Emotion Analysis for Personality Inference from EEG Signals

Abstract

ÑThe stable relationship between personality and EEG ensures the feasibility of personality inference from brain activities. In this paper, we recognize an individualÕs personality traits by analyzing brain waves when he or she watches emotional materials. Thirty-seven participants took part in this study and watched 7 standardized film clips that characterize real-life emotional experiences and target seven discrete emotions. Features extracted from EEG signals and subjective ratings enter the SVM classifier as inputs to predict five dimensions of personality traits. Our model achieves better classification performance for Extraversion (81.08%), Agreeableness (86.11%), and Conscientiousness (80.56%) when positive emotions are elicited than negative ones, higher classification accuracies for Neuroticism (78.38-81.08%) when negative emotions, except disgust, are evoked than positive emotions, and the highest classification accuracy for Openness (83.78%) when a disgusting film clip is presented. Additionally, the introduction of features from subjective ratings increases not only classification accuracy in all five personality traits (ranging from 0.43% for Conscientiousness to 6.3% for Neuroticism) but also the discriminative power of the classification accuracies between five personality traits in each category of emotion. These results demonstrate the advantage of personality inference from EEG signals over state-of-the-art explicit behavioral indicators in terms of classification accuracy.

Frontal EEG Asymmetry and Middle Line Power Difference in Discrete Emotions

“…The remaining features corresponding to nonzero elements in w i can be ranked by the absolute value of these nonzero elements and were selected for recognizing tenderness and amusement. The selected features were entered as inputs to a support vector machine (SVM) classifier, which seeks to separate different classes of examples apart with maximum margin for the hyperplane (Liu et al, 2017 ; Zhao et al, 2017 ). The leave-one-subject-out subject-independent cross validation was utilized, and the average classification accuracy was computed.…”

Guest Editorial: Apparent Personality Analysis

“…From this review, it is clear that the special issue achieved its goal of capturing a snapshot of cutting edge research and applications of apparent personality analysis with automatic methods. Techniques devoted to the prediction of personality traits were presented in most articles [6], [7], [8], [11], [15]. However, each paper focused in a very specific aspect, including, mixing both, learned and handcrafted features [6]; proposing explainability mechanisms [7]; learning from biological signals [15] and affective states [8].…”

“…An interesting study on the use of electroencephalography (EEG) signals for predicting personality traits is presented by G. Zhao et al in their article "Emotion Analysis for Personality Inference from EEG Signals" [15]. The authors designed an experiment involving more than 40 subjects.…”

MBTI Based Personality Prediction o f a User Based o n Their Writing o n Social Media

“…Backpropagation algorithm, which utilized root mean squared error (RMSE) as the metric of quality, was chosen as the training algorithm for the models. Guozhen Zhao et al [4] modelled a paper which aims at recognizing personality traits by analyzing brain waves. Model classifies according to the big five frame work.…”