Application of Extreme Gradient Boosting Ensemble Model for Sleep Quality Prediction on Personalized Wearable Device Data
In this research article, we have applied the XGBoost Ensemble-based supervised machine learning model for sleep quality prediction of an individual person based on the data collected from wearable devices. We have also implemented the most popular supervised machine learning-based model like the random forest, SVM, KNN, Perceptron, SGD, and other models for comparison of the results obtained. From this study, we have found that an extreme gradient boosting based model outperforms other models and achieves the highest precision of 84 % for class 0 data. We have also presented the importance score of the various features used in this experiment for the XGBoost model. We find from this analysis that the supervised machine learning-based models perform well while the data volume is not very high and the features are very small. The models provide a significant amount of accuracy. The models provide accuracy considerable. The highest accuracy is around 73 percent, although the collection of data is not that sufficient.
Keywords: Sleep quality, Wearable Devices, XGBoost, Supervised machine learning.