Urban traffic prediction using metrological data with fuzzy logic, long short-term memory (LSTM), and decision trees (DTs)

In Kuwait, the transport sector is facing a daily traffic congestion pandemic. The traffic congestion is significantly influencing the economy and obstructing the development and production in the country. Extreme weather conditions also affect the roads’ traffic, causing considerable hazards to the transportation system. Precipitation, temperature, wind speed, and visibility are the principal weather variables influencing Kuwait’s traffic. Two selected roads were studied and analyzed to conduct this research. It includes comprehensive deep learning to investigate and analyze the correlation between the weather variables and their impact on Kuwait’s roads and traffic congestion. Heat maps were used as a qualitative measure of the output variables for a better understanding of the data. Three machine learning approaches were selected: fuzzy logic, long short-term memory (LSTM), and decision trees. They were implemented on a dataset from Kuwait Control and Meteorological Center for the year of 2018. Moreover, a validation test of unseen data was implemented to verify the outputs of machine learning. The results indicated that traffic congestion is evidently associated with the weather variables and that the temperature is the essential variable. Machine learning was developed to predict traffic congestion using weather data. The three presented models have demonstrated an overall good performance accuracy in classifying the data based on input features. However, the LSTM has proven to have the best results of the three models.