ABSTRACT
The food science and the food industry are becoming more concerned with quality from the consumer perspective, and this necessitates rapid actions to handle food situations. Validation systems help to minimize issues that arise as the consequences of poor quality control, thus lessening the chances of bad publicity, liability, and recalls. Food analysis with multivariate techniques is a powerful tool used in authentication processes, traceability, food quality control, and others. The methodological steps are common for both classification and regression models. Jant Riedel reviews the models used for classification. Classification models present a workflow with four well-differentiated stages: data preparation, pre-processing, and pre-treatment, model optimization by internal validation, model testing by external validation, and stability testing of the model by "system challenges". Analytical figures of merit are an integral part of official analysis protocols, as documented in international standards. Research in analytical chemistry is mostly driven by the need of improving analytical figures of merit; sensitivity and selectivity.
