Density Based Mode Within Radius and Its Application to EEG based Emotion Recognition

Autor/innen

  • Parampuneet Thind Pace University
  • Vaibhav Katturu Pace University
  • Teryn Cha
  • Sung-Hyuk Cha Pace University

DOI:

https://doi.org/10.32473/flairs.v34i1.128558

Schlagworte:

mode, EEG, Emotion

Abstract

Statistical parameters such as center points are widely used to represent patterns in machine learning. There are three popular center points: mean, median, and mode. Mean and median are pervasively used as features to represent patterns, but mode is not because of its weakness. One of the major problems in the conventional definition of mode occurs when the distribution is multimodal with tied multiple maximum points. Hence, in this paper we propose a modified version of mode, which utilizes the density within a radius. The effectiveness of proposed mode is demonstrated in the emotion recognition application using the Electroencephalogram (EEG). An experimental result of amalgamating statistical parameters with a new modified parameter suggests the superiority of the proposed statistical parameter over the traditional measures.

 

Downloads

Veröffentlicht

2021-04-18

Zitationsvorschlag

Thind, P., Katturu, V., Cha, T., & Cha, S.-H. (2021). Density Based Mode Within Radius and Its Application to EEG based Emotion Recognition. The International FLAIRS Conference Proceedings, 34. https://doi.org/10.32473/flairs.v34i1.128558

Ausgabe

Rubrik

Main Track Proceedings