Portal de Periódicos da CAPES

Using Ensemble Machine Learning Methods to Forecast Particulate Matter (PM2.5) in Bangkok, Thailand

2022; Springer Science+Business Media; Linguagem: Inglês

10.1007/978-3-031-20992-5_18

1611-3349

Patchanok Srisuradetchai, Wararit Panichkitkosolkul,

Forecasting Techniques and Applications

Many areas of Bangkok and its environs are currently blanketed with fine dust with dangerous levels of PM2.5. High levels of PM2.5 have a negative impact on human health. In this study, support vector regression, begged regression tree, random forest, gradient boosted models, neural networks, neural networks autoregressive, seasonal autoregressive moving average with exogenous covariates, k-nearest neighbor, Bayesian additive model, Prophet, and general additive models are used to anticipate PM2.5. The usefulness of adopting an ensemble model for forecasting is investigated. A thorough evaluation of standalone algorithms and ensemble techniques was performed using the root-mean-square error, mean absolute error, and Pearson correlation coefficient. According to the results, hybrid models are effective in the forecasting of PM2.5 concentrations.