Is There a Best Model?

Mathematical models are extensively used to support§decision-making in many disciplines. Nevertheless§there are not any clear standard guidelines to assess§models performance. This significantly affects model§selection processes, which aim to determine the 'best§model', among several possible candidates. Model§performance is often measured by the accuracy with§which models predictions fit independent§observations, however this only assesses a single§aspect of a model. The objective of this book is to§introduce a multi-aspect comparison approach which§intends to provide a comprehensive assessment of§models capabilities. The innovative aspect of this§approach is to create a relationship among four§conventional tests, i.e. uncertainty and sensitivity§analysis, goodness-of-fit prediction-observations,§model complexity and level of details, in order to§provide a reliable estimation of the differences§between constructed and conceptual models. This book§can be of particular interest to modellers and§students interested in having an overview of model§testing as well as scientists willing to apply§radioecological models to predict the contamination§of food products.