Multivariate Analysis

Many processes, both technical and social, depend on several known, and unknown, variables and have complex interactions between them. Multivariable methods are a collection of statistical techniques which pursuit knowledge of a process by analyzing more than one statistical variable simultaneously.

SINTEF offers expertise in several different methods for statistical analysis of multivariable problems like Cluster analysis, Factor analysis, Regression analysis, Principal component analysis, Discriminant analysis and Neural networks.

All these methods are used to identify and model linear and non-linear structures in observation data.

Typical application areas can be:

• Gas measurement. Gas concentration or identification is the basis for a number of process control or safety applications. SINTEF ICT has wide experience in gas sensing. This activity is two-fold: on one hand we focus on high performance gas sensing, on the other the focus is on practical (often low cost) single gas sensors.
• Food quality inspection. Food production faces a number of product quality issues. First, production based on the actual quality of the raw materials will not only help optimise the value of the raw materials, but also enable products with well-defined qualities, e.g. in recipe optimisation. Second, documented product quality enables differentiation of products into different markets, e.g. fatter fish obtains a higher price in Japan than in Europe. Further, increased consumer awareness results in stronger demand on producers on documentation and traceability.
• Real time image processing
• Pattern recognition
• Instrument calibration
• Smart sensors
• Data analysis
• Spectral analysis
• Process modeling

Two collaborating departments within SINTEF develop applications within this special field, Department of Optical Measurement Systems and Data Analysis (our department) and Department of Applied Mathematics. SINTEF is also working in close cooperation with MATFORSK in particular within food inspection.