Approximate Bayesian Computation forForecasting in Hydrological models

Research output: Contribution to conferenceOther

Abstract

Approximate Bayesian Computation (ABC) is a statistical tool for handlingparameter inference in a range of challenging statistical problems, mostlycharacterized by an intractable likelihood function. In this paper, we focus on theapplication of ABC to hydrological models, not as a tool for parametric inference,but as a mechanism for generating probabilistic forecasts. This mechanism is referredas Approximate Bayesian Forecasting (ABF). The abcd water balance modelis applied to a case study on Aipe river basin in Columbia to demonstrate the applicabilityof ABF. The predictivity of the ABF is compared with the predictivity of theMCMC algorithm. The results show that the ABF method as similar performanceas the MCMC algorithm in terms of forecasting. Despite the latter is a very flexibletool and it usually gives better parameter estimates it needs a tractable likelihood
Original languageEnglish
Pages777-782
Number of pages6
Publication statusPublished - 2018

Fingerprint Dive into the research topics of 'Approximate Bayesian Computation forForecasting in Hydrological models'. Together they form a unique fingerprint.

  • Cite this