Pollutant emissions are considered to be a major threat to air quality and human health in urban areas. Therefore, accurate modeling and assessment tools are required. In this study, a model was done by the integration of machine learning algorithms and a geographic information system model. This model included the optimization of the support vector regression model by using the principal component analysis algorithm. Then, the integration of the regression model with spatial analysis modeling via a grid (100 x 100 m) was done in order to generate prediction maps during holidays and workdays in the daytime and at nighttime in a highly congested area in Baghdad city, Iraq. The data used in this study categorized into two categories. The first category is the data acquired through field surveying that includes temperature, humidity, wind speed, wind direction, and traffic flow data (e.g., the number of light and heavy vehicles), as well as carbon monoxide samples by using mobile equipment. The second category is the information derived from geographic information system data, such as land use, road network, and building height. The accuracy of the proposed model is 81%, and the lowest value of root mean square error was 0.067 ppm. The integration between air pollution models and geographic information system techniques could be a promising tool for urban air quality assessment and urban planning. These tools effectively utilized by stakeholders and decision-makers to outline proper plans and strategies to mitigate air pollutants in urban areas.
Urban Air Quality Assessment Using Integrated Artificial Intelligence Algorithms and Geographic Information System Modeling in a Highly Congested Area, Iraq
