Boosting framework via clinical monitoring data to predict the depth of anesthesia.

Background: Prediction of the depth of anesthesia is a difficult job in the biomedical field.

Objective: This study aimed to build a boosting-based prediction model to predict the depth of anesthesia based on four clinical monitoring data.

Methods: Boosting is a framework algorithm that is used to train a series of weak learners into strong learners by assigning different weights according to their classification accuracy. The input of the boosting-based prediction model included four types of clinical monitoring data: electromyography, end-tidal carbon dioxide partial pressure, remifentanil dosage, and flow rate. The output was the depth of anesthesia.

Results: The boosting framework model built in this study achieved higher prediction accuracy and a lower discrete degree in predicting the depth of anesthesia compared with the DT-, KNN-, and SVM-based models.

Conclusions: The boosting framework was used to set up a prediction model to predict the depth of anesthesia based on four clinical monitoring data. In the experiments, the boosting framework model of this study achieved higher prediction accuracy and a lower discrete degree. This model will be useful in predicting the depth of anesthesia.