A. P. Palacios, J. M. Marín Diazaraque, M. P. Wiper
Bacterial growth models are commonly used in food safety for the prediction of microbial safety and the shelf life of perishable foods. Statistical methods for the analysis of growth curves have been widely studied, nevertheless many challenging problems remain. Bacterial growth is observed in Petri dish experiments and the experiment can be replicated several times under the same conditions. Even when the conditions are the same, the observed curves are different, reflecting the intrinsic variability of the growth process. The main idea developed in this paper is to introduce stochastic variability into the deterministic Gompertz growth process by subordination. The resulting stochastic growth model is monotonically nondecreasing and its mean trajectory follows the classical Gompertz equation. We also show how the model can be fitted using a Bayesian approach and we illustrate the methods using real data from a bacterial growth experiment.
Palabras clave: stochastic growth models, bayesian inference, subordination, bacterial growth
Programado
MC3 Bioestadística 1
17 de abril de 2012 12:00
Sala Londres
