Multivariate Analysis and Chemometrics
Insight & Understanding
These procedures are designed to extract useful information from large or complex data sets. Multivariate methods have broad application in many aspects of practical food science including microbiology, chemistry and engineering.
Experiment Design
Statistical experiment design techniques use experimental resources (time, materials, equipment) efficiently to collect data for development of models.
Modeling and Simulation
Models can be developed using regression or other methods and used to find true optima in quality or cost, or to seek acceptable compromises in performance of several factors. These techniques have application in analytical method development and product and process optimization. They are particularly useful for generating process scale-up data, and often minimize surprises during transfer to industrial scale operation.
Composition/Property Relationships
Relationships between product constituents or ingredients and properties can be discerned, and in many cases controlled or optimized. The properties may include sensory (flavor, texture, turbidity, perceived color, etc.) or physical (firmness, light scattering, foam, etc.) properties or cost.
Structure/Function Relationships
Improving understanding of relationships between the structure of molecules (i.e., amino acid composition of peptides) and their biological (flavor) or physical behavior (foam, haze) can lead to improved ingredients and products.
Pattern Recognition
Pattern recognition procedures are useful for discerning which of a number of measurements can discriminate between classes of samples on a desired basis; the results can then be used to classify new samples. Applications include the identification of samples as to cultivar or growing area, or detection of adulteration.
Contact:
Dr. Karl J. Siebert - (Tel: 315-787-2299)
Email: Karl_Siebert@Cornell.edu