Abstract | ||
---|---|---|
When attempting to recognize mental stress using heart rate variability (HRV), single classification models tend to have lower accuracy in detecting different stress levels accurately, and are likely to lead to over fitting, therefore affecting the accuracy of stress recognition. This study employed the ensemble learning method of random forests (RF) and proposed a method to recognize stress by using HRV. By analyzing the short-term (120-180 sec) electrocardiography (ECG) data of the subjects during a stress-inductive video game, we extracted their HRV readings using a time-domain method, frequency-domain method, and non-linear method. Next we constructed a stress recognition model based on the RF technique, which was able to identify low, medium, and high level of stress. Then the model was applied to 200 groups of stress level data collected from the 10 subjects. The results showed that, compared to traditional k-nearest neighbor (KNN) and logistic regression (LR) methods, the RF model could be used to automatically detect and identify stress of different levels with a higher level of accuracy, and with 90% accuracy in recognizing higher levels of stress. |
Year | DOI | Venue |
---|---|---|
2016 | 10.3233/978-1-61499-722-1-444 | Frontiers in Artificial Intelligence and Applications |
Keywords | Field | DocType |
heart rate variability,stress recognition,logistic regression,random forests | Forestry,Random forest,Geography | Conference |
Volume | ISSN | Citations |
293 | 0922-6389 | 0 |
PageRank | References | Authors |
0.34 | 0 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Gang Zheng | 1 | 109 | 19.51 |
Yan-Hui Chen | 2 | 0 | 0.34 |
Min Dai | 3 | 1 | 1.71 |