Title :
A Convex Combination of Models for Predicting Road Traffic
Author :
Bellosta, Carlos J Gil
Author_Institution :
Datanalytics, Madrid, Spain
Abstract :
This paper describes an approach to the road traffic prediction problem in Warsaw in the context of a data mining competition that is part of the IEEE ICDM 2010. A solution based on a convex combination of models mining different wells of information within the data is described. Such convex combination allows the final model compensate highly uncorrelated errors from the different underlying models and to achieve higher prediction accuracy.
Keywords :
data mining; road traffic; traffic control; IEEE ICDM 2010; convex combination; data mining competition; model compensate; prediction accuracy; road traffic prediction; uncorrelated errors; underlying models; c control; data mining; forecasting; predictive modeling; road traffic;
Conference_Titel :
Data Mining Workshops (ICDMW), 2010 IEEE International Conference on
Conference_Location :
Sydney, NSW
Print_ISBN :
978-1-4244-9244-2
Electronic_ISBN :
978-0-7695-4257-7
DOI :
10.1109/ICDMW.2010.23
