Data-driven one day ahead air Quality forecast for NO2 and PM2.5 pollutants

Abstract

Air pollution poses a major threat to our health and safety. Common air pollutants include nitrogen dioxide (NO2) and particulate matter 2.5 (PM2.5), also known as fine suspended particulates (FSP), which are emitted by motor vehicles and power plants. The concentration levels of these pollutants vary from day to day and can affect our daily activities and our health. Due to the detrimental effects of air pollution, it is becoming increasingly important to be able to predict air quality. This project focused on air pollution caused by NO2 and PM2.5 pollutants in Hong Kong and how machine learning techniques can be utilised to predict the concentration levels of these pollutants in the following day. Specifically, tree-based regression techniques, including Gradient Boosted Regression Trees, were explored and their performances were then evaluated using metrics such as Mean Absolute Error and Root Mean Square Error.