Electronic Theses and Dissertations

Identifier

2627

Date

2016

Document Type

Thesis

Degree Name

Master of Science

Major

Electrical and Computer Engr

Concentration

Computer Engineering

Committee Chair

Mohammed Yeasin

Committee Member

Eddie Jacobs

Committee Member

Madhusudhanan Balasubramanian

Abstract

Context-specific interpretation and description of emotions and affective states are critical in designing assistive solutions and affect enabled software agents. Despite the limitations, four-dimensional affective space (Valence, Arousal, Dominance, and Like) are adequate for practical applications. In this thesis, we present a systematic way of analyzing and modeling continuous four-dimensional affective space from EEG recording. Main findings are (but are not limited to): (i) determining functional brain areas and frequency bands that are relevant in modeling affective space, (ii) select a set of most relevant features for each dimensions of the 4-D affective space using recursive feature elimination (RFE) and (iii) model continuous four-dimensional affective space using support vector regression (SVR). We performed empirical analyses to evaluate the performance of the model in predicting continuous affective space. It was observed that the prediction accuracy of the final model is significantly better than human judgment evaluated on same data set.

Comments

Data is provided by the student.

Library Comment

Dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.

Download

Share

COinS