Highlights Travel speed forecasting using non-parametric models. Outperforms standard deep learning architectures such as LSTMs, stacked autoencoders, MLP, and other machine learning models. Accounting for lifecycle costs of models, simpler approaches remain competitive.
Abstract The approach taken by the second place winner of the TRANSFOR prediction challenge is presented. The challenge involves forecasting travel speeds on two arterial links in Xi’an City in China for two five hour periods on a single day. Travel speeds are measured from trajectory information on probe vehicles from a fleet of vehicles for a large sub-area of the city. After experimenting with several deep learning methods, we settle on a simple non-parametric kernel regression approach. The method, borrowed from previous work in fixed route transit predictions, formalizes the intuition that in urban systems most failure patterns are recurrent. Our choice is supported by test results where the method outperformed all evaluated neural architectures. The results suggest simple methods are very competitive, particularly considering the high lifecycle cost of deep learning models.
Keep it simple stupid! A non-parametric kernel regression approach to forecast travel speeds
Tramsportation Research, Part C: Emerging Technologies ; 110 ; 269-274
2019-11-20
6 pages
Article (Journal)
Electronic Resource
English
Forum - Why "Keep It Simple Stupid" employee training methods are, well, kind of stupid.
Online Contents | 2003
|Keep it simple, stupid: Simplicity and versatility remain the key aims of grab suppliers
British Library Online Contents | 1998
Online Contents | 2006