Mauricio Moreno-Zambrano

• In this thesis, a series of mathematical models are presented to bring a quantitative exploration of the mechanistic features of the cocoa bean fermentation process with the use of Ordinary Differential Equation Systems. In this sense, several sources of data have been used in order to fit these mathematical conceptualizations by a Bayesian framework to solve the parameter estimation problem. A baseline model is proposed, assessed and discussed in terms of its biological plausibility, and an interpretation of its got parameter estimates is introduced as an indirect indicative of differences between trials’ features. Therefore, I present a deeper analysis of model iterations based on the baseline with the purpose of accomplishing a wider exploration of five hypothesized mechanisms, e.g., over-oxidation of acetic acid and consumption of fructose by lactic acid bacteria. In that way, their likeliness of occurrence is determined by their overall success on fitting several data gathered from 23 different fermentation trials. Also, an analysis of obtained parameter estimates as classifying fermentation features is discussed. Finally, the effect of temperature on kinetic modeling of cocoa bean fermentation is addressed by the use of Arrhenius terms and discussed in terms of gains in interpretation and biological plausibility of the parameter estimates and its effect on model accuracy. The findings in this thesis provide new insights into the understanding of the complex process of cocoa bean fermentation by assessing candidate mechanisms, and interpreting parameter estimates from a biological point of view towards their use as an addition to chemical fingerprinting methods for classifying features.