A copula-based approach for model bias characterization was previously proposed [18] aiming at improving prediction accuracy compared to other model characterization approaches such as regression and Gaussian Process. This paper proposes an adaptive copula-based approach for model bias identification to enhance the available methodology. The main idea is to use cluster analysis to preprocess data, then apply the copula-based approach using information from each cluster. The final prediction accumulates predictions obtained from each cluster. Two case studies will be used to demonstrate the superiority of the adaptive copula-based approach over its predecessor.
An Adaptive Copula-Based Approach for Model Bias Characterization
Sae Int. J. Mater. Manf
Sae International Journal of Materials and Manufacturing
SAE 2015 World Congress & Exhibition ; 2015
Sae International Journal of Materials and Manufacturing ; 8 , 2 ; 315-321
14.04.2015
7 pages
Aufsatz (Konferenz)
Englisch
An adaptive copula-based approach for model bias characterization
Kraftfahrwesen | 2015
|A Copula-Based Approach for Model Bias Characterization
British Library Conference Proceedings | 2014
|A Copula-Based Approach for Model Bias Characterization
SAE Technical Papers | 2014
|Multinomial Logit model based on Gumbel Copula
British Library Online Contents | 2009
|A Multivariate Copula-Based Macro-Level Crash Count Model
Transportation Research Record | 2018
|