Forecasting Air Quality Index in Thailand Using Ensemble Method

Air pollution is one of the most serious problems in many regions of the world. Thailand also has had to face this trouble unavoidably, especially in the northern region of Thailand, the area that has been highly contaminated by air pollution for so many years. In this paper, an ensemble method was introduced to forecast the level of air quality index (AQI) in the northern part of Thailand. The ensemble method, in this study, is a technique gaining the results from majority vote of outputs of three classification models—k-nearest neighbors, random forest, and support vector machine. The proposed model compared the voted accuracy with the accuracies of existing classification models. It made use of the 2018 - 2021 data from seven stations in four provinces of Northern Thailand. In the end, the proposed model yielded 99.68% - 99.84% accuracy rate on average higher than most of the performance of the other comparative models.