Paper
Mental Stress Level Classification Using Eigenvector Features and Principal Component Analysis
- Authors:
- Sazali Yaacob; Saidatul Ardeenawatie; Paul Murugesa Pandiyan
- Abstract
- This paper presented the efficiency of eigenvector features namely modified covariance to extract the mental stress features. In this work, we have investigated the possibility of applying statistical features and principal component analysis (PCA) to obtain the most informative features that can represent the entire dataset. Electroencephalography signals (EEG) were collected from ten subjects (6 males and 4 females) in a controlled environment using Mental Arithmetic Test (MAT) stimuli. In preprocessing stage, acquired EEG signals were filtered into four frequency bands namely; Delta (0.5 - 4 Hz), Theta (4 - 7 Hz), Alpha (7 - 13 Hz) and Beta (13 – 30 Hz) using elliptic bandpass filter. Extracted features were mapped into corresponding stress level classes; Low (Level 1), Medium (Level 2) and High (Level 3) using K-Nearest Neighbours (KNN) classifier. The classification accuracy of 98% confirmed that the proposed Brain Computer Interfaces (BCI) system combining eigenvector features and PCA have potential in classifying the mental stress level.
- Keywords
- Electroencephalography Signal (EEG); K-Nearest Neighbours (KNN); Mental Stress; Modified Covariance; Principal Component Analysis (PCA)
- StartPage
- 254
- EndPage
- 261
- Doi