Why Random Forest Is Not the Best Model?

The latest value model is most important, at least to me. It is a naive model where the traffic value that is predicted is actually the latest one that was seen by the model. The traffic time serios do not change too much between following samples. So i would recommend to u a random forest, for example. But if you use a complex deep learning with a tong of internal liers, you are not going to excel that performance.