Forecasting CO pollutant concentration of Tabriz city air using artificial neural network and adaptive neuro-fuzzy inference system and its impact on sustainable development of urban

Abstract

Tabriz as one of the major industrial cities of Iran is not immune to air pollution and spends many days air-polluted each year. Since one of the goals of sustainable development is to achieve clean and good air for all segments of society and to attract clean air from the citizens of a city. In this study, it was attempted to present an efficient model for predicting CO pollutant concentrations using artificial neural network (ANN) and adaptive neural-fuzzy inference system (ANFIS). Air quality monitoring and developing efficient air pollution models can be suitable in providing requirements for sustainable development goals. For modelling, the meteorological and pollutant data were first obtained from the Meteorological and Environmental Agency of Tabriz city. The model inputs were temperature, wind speed, humidity and contaminant concentration daily hour index and weekly day index. The results of the study at validation step yielded 0.82 and 0.63 in terms of determination coefficient for ANN models and ANFIS models. It was also observed that the Ensemble method worked even better than the single 2 methods. Keywords Air pollution • Artificial neural network • Adaptive fuzzy neural inference system • CO • TabrizThis article is a part of the Topical collection in Environmental Earth Sciences on "Water Problems in E. Mediterranean Countries" guest edited by H. Gökçeku, D. Orhon, V. Nourania, and S. Sozen.

Cause Identification and Coupling Relationship Analysis of Urban Problems: A Case Study of Poor Parking Convenience

“…However, when assessing factor correlations, subjective and qualitative evaluations of factors are often required. Real systems usually involve inaccurate and uncertain information, leading to the incorporation of grey theory and fuzzy logic for optimization [40]. When applying fuzzy logic, the selection of an appropriate membership function for the research issue is crucial, whereas grey methods demonstrate distinct advantages when dealing with uncertainties arising from small samples, inaccuracies, and incomplete information [41].…”

Regional VOCs Gathering Situation Intelligent Sensing Method Based on Spatial-Temporal Feature Selection

“…Zhu et al developed a support vector regression (SVR) model combined with the cuckoo search algorithm (CS) and the grey wolf optimisation algorithm (GWO) to model high and low frequency sequences to predict NO 2 or SO 2 in central China [33]. Nourani et al produced an effective model using ANN and ANFIS to predict CO pollutant concentrations [34]. Wong et al combined [40].…”