Project Awarded under : Collaborative Research Scheme(CRS) of TEQIP-III through ASTU
About the project:Air pollutants in and around an industrialized and densely populated city is asubject of study. It is a big concern to the human health conditions and the individualsare also susceptible to different health hazards if the presence of thesepollutants crosses permissible levels. Respirable Suspended Particulate Matters(RSPM) have a great impact on human health.In this project, we compare ARIMA and machine learningalgorithms – Linear Regression, Neural Network Regression, Lasso ElasticNet, DecisionForest Regression, Extra Trees Regression, Decision Tree Ada-Boost, XGBoost.Our primary focus is to find how well each model forecasts the seventeen yearsdata collected from Pollution Control Board Assam (PCBA). PCBA is monitoringair through nationwide programs. This paper aims to analyse the spread of thepollutants in the air focusing mainly on respirable suspended particulate matter(RSPM) and to predict for the most polluted district. We found that data has astrong seasonal structure, the ARIMA model outperforms the machine learningalgorithms though all the methods tend to perform poorly in forecasting becausethere is randomness in the data.
PI: Dr. Th. Shanta Kumar, Associate Professor, Department of CSE