| Title |
Comparison of Classifier for Pain Assessment based on Photoplethysmogram and Machine Learning |
| Authors |
임지연(Ji Yeon Yim) ; 신항식(Hangsik Shin) |
| DOI |
https://doi.org/10.5370/KIEE.2019.68.12.1626 |
| Keywords |
Classifier; Machine Learning; Pain Assessment; Photoplethysmogram |
| Abstract |
This study examines the classification characteristics of various machine learning classifiers for pain assessment using photoplethysmogram.
The presence of pain was assessed using waveform characteristics derived from photoplethysmogram obtained from 73 patients before and after surgery. Classification performance was evaluated using logistic regression, random forest, multilayer perceptron, and 1D convolutional neural network, and was validated with nested kfold cross validation. As a result, pain classification accuracy was highest in order of logistic regression, convolutional neural network, multilayer perceptron, and random forest classifier. In addition, logistic regression, random forest, multilayer perceptron, and convolutional neural network were shown to be robust to overfitting in order. |