Читать книгу Cyberphysical Smart Cities Infrastructures. Optimal Operation and Intelligent Decision Making онлайн

47 страница из 94

To have a safe environment, air pollution prediction in smart cities plays an imperative role. There have been a significant amount of work tried to enhance the prediction using several machine learning algorithms. The usage of machine learning algorithms to make environment safe has increased consistently, stating that how important this prediction would be for smart cities [6]. We examined the most relevant research studies [4, 6, 23, 38] but not limited to these, applying different evaluation based on several metrics. The evaluation of those research leads us to the following common observations: first, the rate of applying of the advanced (DL) machine learning algorithms has proliferated rather than typical machine learning algorithms; second, among the prediction elements for air pollution prediction, PM2.5 is considered as the most popular element; third, the data used for air pollution prediction had already generated hourly rather than daily; and in the final observation, efficient prediction occurs when air‐quality captured data merged with other relevant data of other networks like healthcare network within the smart cities.