Toward forecasting future day air pollutant index in Malaysia

Abstract

The association of air pollution and the magnitude of adverse health effects are receiving close attention from the world. The effects of air pollution were found to be most significant for children, elderly, and patients with preexisting respiratory problems. The existing API forecast system is capable of predicting the air quality based on the pollutant concentrations before critical levels of air pollution are exceeded. However, there is no API forecasting system available in Malaysia that can predict the coming day API readings. This paper aims to propose an API forecast system that utilizes the hourly API in Malaysia to predict the next day API. The proposed solution allows sensitive populations to plan ahead of their daily activities and provide governments with information for public health alerts. We also propose strategies for aggregated-level predictions within the region. Nevertheless, it can be extended across the region, especially in the less economically developed regions across the world. We conduct experiments on the public API dataset to demonstrate the viability of the proposed solution.