Predicting long-term population dynamics with bagging and boosting of process-based models

作者：

Highlights：

• A novel methodology for learning ensembles of process-based models for long-term predictions

• We adapt the methods of bagging and boosting to the task of modeling dynamic systems

• We identify the design decisions for learning such ensembles

• We apply and evaluate the performance of the developed algorithms on modeling tasks of population dynamics in aquatic ecosystems

• Ensembles of PBMs have better predictive performance than a single process-based model.

摘要

•A novel methodology for learning ensembles of process-based models for long-term predictions•We adapt the methods of bagging and boosting to the task of modeling dynamic systems•We identify the design decisions for learning such ensembles•We apply and evaluate the performance of the developed algorithms on modeling tasks of population dynamics in aquatic ecosystems•Ensembles of PBMs have better predictive performance than a single process-based model.

论文关键词：Ensembles,Process-based modeling,Bagging,Boosting,Machine learning,Predictive modeling,Population dynamics

论文评审过程：Received 20 February 2015, Revised 23 June 2015, Accepted 5 July 2015, Available online 18 July 2015, Version of Record 29 August 2015.

论文官网地址：https://doi.org/10.1016/j.eswa.2015.07.004