Detecting concentration of Ozone based on different machine learning methods

Abstract

Hundreds of years ago, the environment was not that polluted; however, as the industrial revolution came, although our life became much easier, many toxic substances were released. These toxic substances result in Ozone pollution. Thus, the topic of the research focuses on predicting the amount of ozone present in the atmosphere in some areas of the world and how different factors, like environmental issues, can affect the pollution of Ozone. Moreover, which factors contribute the most. Through the extensive literature review and retrieval, it raised the question of how different methods of detecting concentrations of Ozone have drawbacks and advantages. In short, the main conclusion is that the simple methods (KNN and Perceptron) we are familiar with are enough to produce good enough results, and though these two simple methods have some disadvantages, there are more advanced methods (GBRT) that make the results even better.