Evaluation of the predictive capacity of five plant growth models for sugar beet

Baey, Charlotte; Li, Song; Cournède, Paul-Henry; Lemaire, Sébastien; Maupas, Fabienne

doi:10.1109/pma.2012.6524809

Cited by 4 publications

(6 citation statements)

References 24 publications

(28 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Provided some basic knowledge in C++, it is very easy to implement models and to use the proposed methods: approximately 20 models of plant growth in interaction with the environment and variants are currently implemented in the platform: mostly in the GreenLab family [21] and variants of the STICS [6], CERES [32], SUNFLO [40] models, for different species. As such, it provides an interesting tool in the process of model comparison and benchmarking in the plant growth community as illustrated in [2]. Even if some specific classes of the platform are specifically derived for plant growth models (environment class, specific plant architectural class, some tools to handle experimental data ...) and the platform is so far exclusively used in the plant growth modeling community, its scope is actually broader and could find applications for other discrete dynamic models.…”

Section: Discussionmentioning

confidence: 99%

“…The data used in this paper for the evaluation of the model were obtained from experiments conducted in 2010 by the the French Institute for Sugar Beet Research (ITB) and presented in [2]. Field experiments took place at La Selve, France, N49…”

Section: Experimental Data For Model Evaluationmentioning

confidence: 99%

“…Corrected AIC for LNAS model with 1 to 6 estimated parameters like the mean-square error of prediction (MSEP) testing the predictive capacity of a model, when two independent data sets are available [64]. An example of use can be found in [2] for the benchmarking of several plant models for sugar beet yield prediction.…”

Section: Model Selectionmentioning

confidence: 99%

See 2 more Smart Citations

Development and Evaluation of Plant Growth Models: Methodology and Implementation in the PYGMALION platform

Cournède

Chen

et al. 2013

Math. Model. Nat. Phenom.

View full text Add to dashboard Cite

Abstract. Mathematical models of plant growth are generally characterized by a large number of interacting processes, a large number of model parameters and costly experimental data acquisition. Such complexities make model parameterization a difficult process. Moreover, there is a large variety of models that coexist in the literature with generally an absence of benchmarking between the different approaches and insufficient model evaluation. In this context, this paper aims at enhancing good modelling practices in the plant growth modeling community and at increasing model design efficiency. It gives an overview of the different steps in modelling and specify them in the case of plant growth models specifically regarding their above mentioned characteristics.Different methods allowing to perform these steps are implemented in a dedicated platform PYGMALION (Plant Growth Model Analysis, Identification and Optimization). Some of these methods are original. The C++ platform proposes a framework in which stochastic or deterministic discrete dynamic models can be implemented, and several efficient methods for sensitivity analysis, uncertainty analysis, parameter estimation, model selection or data assimilation can be used for model design, evaluation or application.Finally, a new model, the LNAS model for sugar beet growth, is presented and serves to illustrate how the different methods in PYGMALION can be used for its parameterization, its evaluation and its application to yield prediction. The model is evaluated from real data and is shown to have interesting predictive capacities when coupled with data assimilation techniques.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Experimental Data For Model Evaluationmentioning

confidence: 99%

See 1 more Smart Citation

Development and Evaluation of Plant Growth Models: Methodology and Implementation in the PYGMALION platform

Cournède

Chen

et al. 2013

Math. Model. Nat. Phenom.

View full text Add to dashboard Cite

show abstract

“…With the purpose of testing the performance of the proposed algorithm, we performed simulations based on the true experimental conditions of one experiment conducted by the French institute for Sugar beet research in 2010 (see [2] for a detailed description of the experimental protocol). Here, we present the results of simulations in two different scenarios: (i) the complete dataset is observed, that is, 320 observations (from day 1 to day 160) and (ii) we have only access to 28 observations from this dataset, green leaf and root masses measured at the 14 dates included in O 2010 : 68, 76, 83, 90, 98, 104, 110, 118, 125, 132, 139, 145, 160} , which exactly correspond to the 14 dates of measurement in the 2010 experiment.…”

Section: A Simulation Case Study For a Plant Growth Modelmentioning

confidence: 99%

A Regularized Particle Filter EM Algorithm Based on Gaussian Randomization with an Application to Plant Growth Modeling

Chen

Trevezas

Cournède

2015

Methodol Comput Appl Probab

View full text Add to dashboard Cite

International audienceParameter estimation in complex models arising in real data applications is a topic which still attracts a lot of interest. In this article, we study a specific data and parameter augmentation method which gives us the opportunity to estimate more easily the parameters of the initial model. For this reason, the notion of Gaussian randomization of a model with respect to some of its parameters is introduced. The initial model can be regarded as a submodel of the resulting extended incomplete data model. Under the assumption that the initial model has a unique maximum likelihood estimator (MLE) and that the likelihood function is continuous we prove that the extended model has a unique MLE with common values for the parameters of the MLE which correspond to the initial model. We also prove the reverse direction. Moreover, an appropriate stochastic version of an EM (Expectation-Maximization) algorithm is suggested to make parameter estimation feasible. In particular, we describe how the regularized particle filter of can be used in this frequentist-based approach to perform the Monte Carlo E-step at each iteration of the stochastic EM algorithm. This regularized version is particularly adapted to the framework of Gaussian randomization since the last iterations of the EM algorithm are characterized by low variance in the parameter distributions. A toy example with available analytic solutions, a synthetic example and a real data application with scarce observations to the LNAS (Log-Normal Allocation and Senescence) model of sugar beet growth are presented to highlight some theoretical and practical aspects of the proposed methodology

show abstract

“…The data used for this study were obtained by the French institute for sugar beet research (ITB, Paris, France) in 2006, 2008 and 2010 with slightly different cultivars and in different locations (details of the experimental protocols can be found in [2] and [20]). For the test case, we choose 2010 as the dataset for calibration since more observation points are available compared to the other two datasets (of course, in real applications, the most aged datasets are supposed to be used for the calibration).…”

Section: Experimental Datamentioning

confidence: 99%

Iterative convolution particle filtering for nonlinear parameter estimation and data assimilation with application to crop yield prediction

Chen¹,

Trevezas²,

Cournède³

2013

2013 Proceedings of the Conference on Control and Its Applications

Self Cite

View full text Add to dashboard Cite

The complexity of plant growth models and the scarcity of experimental data make the application of conventional data assimilation techniques rather difficult. In this paper, we use the Convolution Particle Filter (CPF) and an iterative adaptation, the Iterative Convolution Particle Filter (ICPF) for nonlinear parameter estimation. Both methods provide prior distributions in the Bayesian framework for data assimilation. CPF is sequentially used to update state and parameter estimates in order to improve model prediction and to assess the predictive uncertainty. The predictive performances of the two methods are evaluated by an application to the LNAS sugar beet growth model with three sets of real measurements, one used for parameter estimation and the two others used to test the model predictive capacity, both with and without data assimilation. Despite the low accuracy and the scarcity of the early data used for assimilation, the CPFbased data assimilation approach with the prior distribution based on ICPF estimations showed promising predictive capacities and provided robust confidence intervals. The method can therefore be considered as a potential candidate for yield prediction applications in agriculture.

show abstract

Evaluation of the predictive capacity of five plant growth models for sugar beet

Cited by 4 publications

References 24 publications

Development and Evaluation of Plant Growth Models: Methodology and Implementation in the PYGMALION platform

Development and Evaluation of Plant Growth Models: Methodology and Implementation in the PYGMALION platform

A Regularized Particle Filter EM Algorithm Based on Gaussian Randomization with an Application to Plant Growth Modeling

Iterative convolution particle filtering for nonlinear parameter estimation and data assimilation with application to crop yield prediction

Contact Info

Product

Resources

About