Using wavelet transform for features extraction and machine learning algorithms to classify the facial expression by eeg signals
Year: 2022
Type: article
Abstract: The goal was to evaluate the performance of the state-of-the-art algorithms. A secondary goal was to try to improve upon the result of a method that was used in a study similar to the one used in this work. This paper presented the building of multi-state datasets relating to eye behaviors and facial expressions. Signals were recorded and stored by the connection of a channel-less mobile device. Z-score, max-min normalization techniques were used to optimize data. The cross-validation technique divided the data into training/testing segments. The features of the electrical brain signals (delta, theta, alpha and beta band) were analyzed by the Daubechies wavelet transform method. The extracted time and frequency domain features calculate total energy, detailed component energy, approximate component energy, relative energy. Three algorithms, support vector machine, k-nearest neighbor, and ensemble algorithm, were used to develop into 17 models to optimize the classification efficiency of the machine learning algorithms. Parameters of these models were surveyed and optimized to propose a best classification one for the Data-021 dataset. The Subspace ensemble model was proposed because its model efficiency was more than 87,7%.
Author: Tiên Hoàng Thủy Võ
Language: English
Topic: Emotion and Mood Recognition
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4391370607
