Research Article
Using an Ensembled Boosted Model for IoT Time Series Regression
@INPROCEEDINGS{10.1007/978-3-030-94182-6_28, author={Shuai Lin and Kun Zhang and Renkang Geng and Liyao Ma}, title={Using an Ensembled Boosted Model for IoT Time Series Regression}, proceedings={IoT and Big Data Technologies for Health Care. Second EAI International Conference, IoTCare 2021, Virtual Event, October 18-19, 2021, Proceedings, Part II}, proceedings_a={IOTCARE PART 2}, year={2022}, month={6}, keywords={Traffic flow data IoT Time Series Ensemble learning XGBoost}, doi={10.1007/978-3-030-94182-6_28} }- Shuai Lin
Kun Zhang
Renkang Geng
Liyao Ma
Year: 2022
Using an Ensembled Boosted Model for IoT Time Series Regression
IOTCARE PART 2
Springer
DOI: 10.1007/978-3-030-94182-6_28
Abstract
As a typical regression model, time series prediction is a very important part of machine learning. With the development of urban roads and the increase of car ownership, traffic data is more closely related to machine learning. As a common time-series data in our life, traffic flow data has great research value and a wide range of application fields. Compared with the general time series, traffic flow data has larger data volume, stronger volatility, and higher requirements for accuracy and speed of prediction, but traditional algorithms often fail to achieve these goals. With the development of ensemble algorithm, it has outstanding performance in the field of classification and regression. Therefore, we choose XGB algorithm as the core algorithm of this experiment. In this paper, we introduce the working principle of the XGBoost algorithm, the acquisition of the traffic flow data used in the experiment, and the feature extraction of the traffic flow data in detail. Finally, we use the XGBoost algorithm to model and output the prediction results. In addition, we modified some very important parameters in XGBoost, such as iteration model, iteration number, etc., to explore the influence of each parameter on prediction accuracy when the XGBoost algorithm is used to predict traffic flow data.
