Predictive Process-based Modeling of Aquatic Ecosystems
Nina Vidmar, Nikola Simidjievski, Sašo Dzeroski Faculty of Mathematics and Physics, University of Ljubljana, Ljubljana, Slovenia & Jozef Sefan Institute, Ljubljana, Slovenia & Jozef Stefan International Postgraduate School, Ljubljana, Slovenia
Abstract: In this paper, we consider the task of learning interpretable process-based models of dynamic systems. While most case studies have focused on the descriptive aspect of such models, we focus on the predictive aspect. We use multi-year data, considering it as a single consecutive dataset or as several one-year datasets. Additionally, we also investigate the effect of interpolation of sparse data on the learning process. We evaluate and then compare the considered approaches on the task of predictive modeling of phytoplankton dynamics in Lake Zürich.
References:[1] Langley, P. W., Simon, H. A., Bradshaw, G., Zytkow, J. M. (1987). Scientific Discovery: Computational Explorations of the Creative Processes. MA: The MIT Press, Cambridge, MA, USA.
[2] Dzeroski, S., Todorovski, L. (2003). Learning population dynamics models from data and domain knowledge. Ecological Modelling, 170, 129– 140.
[3] Bridewell, W., Langley, P. W., Todorovski, L., Džeroski, S. (2008). Inductive process modeling. Machine Learning, 71, 1–32.
[4] Cerepnalkoski, D., Taškova, K., Todorovski, L., Atanasova, N., Džeroski, S. (2012). The influence of parameter fitting methods on model structure selection in automated modeling of aquatic ecosystems. Ecological Modelling, 245 (0) 136–165.
[5] Taškova, K., Silc, J., Atanasova, N., Dzeroski, S. (2012). Parameter estimation in a nonlinear dynamic model of an aquatic ecosystem with metaheuristic optimization. Ecological Modelling, 226, 36–61.
[6] Atanasova, N., Recknagel, F., Todorovski, L., Dzeroski, S., Kompare, B. (2006a). Computational assemblage of Ordinary Differential Equations for Chlorophyll-a using a lake process equation library and measured data of Lake Kasumigaura. In: Recknagel, F.(Ed.), Ecological Informatics. Springer, 409–427.
[7] Atanasova, N., Todorovski, L., Dzeroski, S., Remec, R., Recknagel, F., Kompare, B. (2006c). Automated modelling of a food web in Lake Bled using measured data and a library of domain knowledge. Ecological Modelling, 194 (1-3), 37–48.
[8] Whigham, P., Recknagel, F. (2001). Predicting Chlorophyll-a in freshwater lakes by hybridising process-based models and
genetic algorithms. Ecological Modelling, 146 (13) 243–251.
[9] Bridewell, W., Asadi, N. B., Langley, P., Todorovski, L. (2005). Reducing overfitting in process model induction. In: Proceedings
of the 22nd International Conference on Machine learning. (ICML ’05). ACM, 81–88.
[10] Simidjievski, N., Todorovski, L., Džeroski, S. (2014). Learning ensembles of population dynamics models and their application
to modelling aquatic ecosystems. Ecological Modelling (In Press).
[11] Todorovski, L., Dzeroski, S. (2007). Integrating domain knowledge in equation discovery. In: Džeroski, S., Todorovski, L.
(Eds.), Computational Discovery of Scientific Knowledge. Vol. 4660 of Lecture Notes in Computer Science. Springer Berlin, 69–
97.
[12] Todorovski, L., Bridewell, W., Shiran, O., Langley, P. W. (2005). Inducing hierarchical process models in dynamic domains.
In: Proceedings of the 20th National Conference on Artificial Intelligence. AAAI Press, Pittsburgh, USA, 892–897.
[13] Durillo, J. J., Nebro, A. J. (2011). jMetal: A Java framework for multi-objective optimization. Advances in Engineering
Software, 42, 760–771.
[14] Storn, R., Price, K. (1997). Differential Evolution – A simple and efficient heuristic for global optimization over continuous
spaces. Journal of Global Optimization, 11 (4) 341–359.
[15] Cohen, S. D., Hindmarsh, A. C. (1996). CVODE, a stiff/nonstiff ODE solver in C. Computers in Physics, 10 (2) 138–143.
(March).
[16] Breiman, L. (1984). Classification and Regression Trees. Chapman & Hall, London, UK.
[17] Dietzel, A., Mieleitner, J., Kardaetz, S., Reichert, P. (2013). Effects of changes in the driving forces on water quality and
plankton dynamics in three swiss lakes long-term simulations with BELAMO. Freshwater Biology 58 (1) 10–35.
[18] Atanasova, N., Todorovski, L., Dzeroski, S., Kompare, B. (2006b). Constructing a library of domain knowledge for automated
modelling of aquatic ecosystems. Ecological Modelling, 194 (13) 14–36.
[19] Wilcoxon, F. (1945). Individual comparisons by ranking methods. Biometrics, 1, 80–83.
[20] Demšar, J. (2006). Statistical comparisons of classifiers over multiple data sets. Journal of Machine Learning Research, 7,
1–30. (December).