Improved ANFIS Model Based on Confinement Optimized PSO to Predict the Air Quality of Delhi, India

This paper proposes an improved version of the adaptive neuro-fuzzy inference system (ANFIS) to forecast the air quality index in New Delhi, India. We have used confinement optimized particle swarm optimizer (COPSO) to improve the performance of ANFIS, where COPSO is an improved version of particle swarm optimizer (PSO). This proposed model has been trained with previous nine years air quality index time series data of Delhi city to forecast the air pollutants such as particulate matter (PM10), sulfur dioxide (SO
2
) and carbon monoxide (CO) as these three air pollutants are found to have severe impact on the air quality of Delhi. We compare the performance of the proposed model (COPSO-ANFIS) with the performance of PSO-ANFIS, polynomial regression and decision tree-based models, where the result shows a significant improvement compared to other approaches.