Fear Level Classification Based on Emotional Dimensions and Machine Learning Techniques.

Oana Bălan,Gabriela Moise,Marius Leordeanu,Florica Moldoveanu,Alin Moldoveanu

doi:10.3390/s19071738

Abstract

There has been steady progress in the field of affective computing over the last two decades that has integrated artificial intelligence techniques in the construction of computational models of emotion. Having, as a purpose, the development of a system for treating phobias that would automatically determine fear levels and adapt exposure intensity based on the user’s current affective state, we propose a comparative study between various machine and deep learning techniques (four deep neural network models, a stochastic configuration network, Support Vector Machine, Linear Discriminant Analysis, Random Forest and k-Nearest Neighbors), with and without feature selection, for recognizing and classifying fear levels based on the electroencephalogram (EEG) and peripheral data from the DEAP (Database for Emotion Analysis using Physiological signals) database. Fear was considered an emotion eliciting low valence, high arousal and low dominance. By dividing the ratings of valence/arousal/dominance emotion dimensions, we propose two paradigms for fear level estimation—the two-level (0—no fear and 1—fear) and the four-level (0—no fear, 1—low fear, 2—medium fear, 3—high fear) paradigms. Although all the methods provide good classification accuracies, the highest F scores have been obtained using the Random Forest Classifier—89.96% and 85.33% for the two-level and four-level fear evaluation modality.

Highlights

Emotion is defined as a conscious mental reaction subjectively experienced and directed towards a specific object, accompanied by physiological and behavioral changes in the body [1]
For the condition where the raw EEG values were included in the model, for the two-level evaluation modality, the highest F scores were obtained using k- Nearest Neighbors (kNN) (86.81% with no feature selection, 84.70% using Fisher selection and 85.54% with Principal Component Analysis (PCA)) and Support Vector Machine (SVM) (60%, using Sequential Forward Selection (SFS) selection)
For the four-level evaluation modality, the highest F scores were obtained using RF (83.55% with no feature selection), kNN (79.22% using the Fisher selection technique and 81.67% using PCA) and SVM and Linear Discriminant Analysis (LDA), both scoring an accuracy of 48% when the SFS algorithm was employed

Summary

Introduction

Emotion is defined as a conscious mental reaction subjectively experienced and directed towards a specific object, accompanied by physiological and behavioral changes in the body [1]. The field of affective computing aims to enhance the interaction between the human and the machines by identifying emotions and designing applications that automatically adapt to these changes. Cognitive behavioral sciences, healthcare and entertainment, affective computing deals with recognizing and modelling human emotions in a way that would improve overall user experience. Some discrete and dimensional models have been proposed and applied throughout the years. The discrete models of affect rely on the existence of a set of basic emotions from which more complex ones derive.

Methods

Results

Discussion

Conclusion