Determining the Effect of Correlation between Asthma/Gross Domestic Product and Air Pollution

Abstract

Air pollution, Asthma, and Gross Domestic Product (GDP) are very important indicators to human life and development and it has been found that air pollution has a big effect on the latter two. In this paper, we find the correlation factor and to what extent air pollution has an effect on those two. For this, we chose 20 American states, handpicked the ones having unique features with respect to pollution levels, asthma cases, GDP numbers, and the datasets for the past 20 years of each state were taken. We chose 6 toxic pollutants, namely PM2.5, Carbon Monoxide, Sulfur Dioxide, PM10, Ozone, and Nitrogen Dioxide with each dataset including daily readings of these pollutants for the past 20 years in each state. The idea behind our model is to use all these data and find the extent to which air pollution is related to the asthma cases and the GDP of a state. For this, we use 4 models, namely Neural Network (NN), Random Forest (RFC), Support Vector Machines (SVM) and K-Nearest Neighbors (KNN). We use metrics like Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and R-Squared to evaluate our results. We observed a positive correlation between rates of asthma and GDP and pollution data. NN gave the best prediction accuracy especially for GDP (Average: 76%) followed closely by SVM. SVM's also had the least MAE while RFC had the least RMSE.