References & Citations
Computer Science > Neural and Evolutionary Computing
Title: Feature Selection on Thermal-stress Dataset
(Submitted on 8 Sep 2021)
Abstract: Physical symptoms caused by high stress commonly happen in our daily lives, leading to the importance of stress recognition systems. This study aims to improve stress classification by selecting appropriate features from Thermal-stress data, ANUstressDB. We explored three different feature selection techniques: correlation analysis, magnitude measure, and genetic algorithm. Support Vector Machine (SVM) and Artificial Neural Network (ANN) models were involved in measuring these three algorithms. Our result indicates that the genetic algorithm combined with ANNs can improve the prediction accuracy by 19.1% compared to the baseline. Moreover, the magnitude measure performed best among the three feature selection algorithms regarding the balance of computation time and performance. These findings are likely to improve the accuracy of current stress recognition systems.
Link back to: arXiv, form interface, contact.