The Loughborough researchers developed machine learning system to predict air pollution index for 1, 3 and 6 hours ahead in advance. The developed machine learning system can automatically learn the features that affect air pollution index and use them to establish predictive models and generate prediction. Explainable and interpretable machine leaning models are developed which could lead to better understanding how complex conditions such as the weather, seasonal and environmental factors would impact on air pollution levels Benefit from an uncertainty analysis process, the system can also quantify the uncertainty in prediction, model structure, parameters.