Air Quality Predicting System for Colombo City using Machine Learning Approaches

Abstract

Abstract Air pollution is one of the biggest threats to the environment and human beings. Because of the meteorological and traffic factors, the burning of fossil fuels, industrial activities, power plant emissions acts as major effects for air pollution. Therefore, the governments of the developing countries like Sri Lanka are majorly focused on the effects of air pollution and they create the rules & regulations to minimize the level of air pollution. The main purpose of this study is to design a Machine Learning approach to predict air pollution status and levels in Colombo city by analyzing the previous dataset of PM2.5 air pollutants. This paper presents, how previous researches predict the air quality level using different types of technologies and data collection methods used to analyze the air quality. And also, it demonstrates the design and implementation of an air quality predicting system, named as Air Quality Predicting System for Colombo City using Machine Learning Approaches. A simple Linear Regression-based supervised machine learning algorithm is using for the predicting process and it gives 8.578 average Root Mean Squared Error (RMSE) value with higher accuracy. This system will implement in both web and mobile platform and it will provide a better user experience. In Sri Lanka, there is no way to predict the air quality based on the above scenario. Most of the researchers have used PM2.5 air pollutant concentration levels as the main feature of their approaches due to the higher relationship to the Air Quality Index value. And also those researches are mostly based on supervised machine learning algorithms like Linear regression, FFNN, & SVM algorithms.