A novel concentration prediction technique of carbon monoxide (CO) based on beluga whale optimization-extreme gradient boosting (BWO-XGBoost)

Background

Carbon monoxide (CO), as a toxic gas, poses serious threats to human health and the ecosystem. Monitoring CO concentration is imperative. However, current algorithms used for CO concentration detection suffer from low accuracy due to limitations in data processing and model training. These algorithms fail to adequately consider the complexity and non-linear relationships within CO concentration data, necessitating the search for a more effective and precise approach.

Methods

In this study, we first introduce an MI-RF feature selection algorithm combining Mutual Information (MI) with Random Forest (RF) to extract key features. Subsequently, we introduce BWO-XGBoost, which combines Beluga Whale Optimization (BWO) and Extreme Gradient Boosting (XGBoost) to achieve higher prediction accuracy.

Significant Findings

We compare it with traditional models such as K-Nearest Neighbors (KNN), Gradient Boosting Decision Tree (GBDT), Support Vector Regression (SVR), and XGBoost. Experimental results demonstrate that the proposed BWO-XGBoost exhibits superior performance in terms of fitting and prediction accuracy.