Chemometrics, Bioinformatics

The need to consider variability due to raw materials, seasonality, agricultural practices, and food processing, that are aspects which all play a role in authenticity tasks, justifies the need for chemometrics methods. This chapter presents a few basic chemometrics methods, such as exploratory data analysis, multiway data analysis, and data fusion. New chemometrics methods are continuously being updated and improved upon, two main distinctive characteristics are required: data exploration and graphical representation; and deep model validation through all steps of data processing. This also explains why chemometrics tools based on latent variables, for example, principal component analysis (PCA), soft independent modeling of class analogies (SIMCA), partial least squares (PLS), and PLS discriminant analysis (PLSDA), are still so popular and powerful. Nowadays, there is an established set of chemometric multiway methods and algorithms. The chapter mentions those that can serve the purposes of exploratory data analysis and classification, the tasks most frequently encountered in food authentication.