1.WHAT A NEW ML DO FOR TRAFFIC?
A new machine learning model can predict traffic in different urban zones. to do this, researchers at the complexity science hub used data from a leading italian car-sharing company as a proxy for all urban transportation. for example, understanding how different urban zones interact can help avoid traffic congestion and enable targeted political responses such as regional expansion of public transportation.
2.HOW IT IS HELPFUL?
As populations grow in urban areas, this knowledge will help policy makers design and implement effective transport policies and comprehensive urban planning. if people commute from one zone to another for a specific reason, you can offer services that complement this interaction. on the other hand, if models show little activity in a particular place, policy makers can use this knowledge to invest in structures to change that.
3.WHY THE NEW MODEL IS REALLY INTERPRETABLE?
There are already many models designed to predict traffic behavior in cities, but the majority of predictive models are not fully interpretable with aggregated data. some structures in the model connect two zones, but cannot be interpreted as an interaction. this limits our understanding of the underlying mechanisms that govern citizens’ daily lives. the new model is fully interpretable as only minimal constraints are considered and all parameters represent real interactions.