A new approach of combination of near-infrared (NIR) spectroscopy and refractometry was developed in this work to determine the concentration of alcohol and real extract in various beer samples. A partial least-squares (PLS) regression, as multivariate calibration method, was used to evaluate the correlation between the data of spectroscopy/refractometry and alcohol/extract concentration. This multivariate combination of spectroscopy and refractometry enhanced the precision in the determination of alcohol, compared to single spectroscopy measurements, due to the effect of high extract concentration on the spectral data, especially of nonalcoholic beer samples. For NIR calibration, two mathematical pretreatments (first-order derivation and linear baseline correction) were applied to eliminate light scattering effects. A sample grouping of the refractometry data was also applied to increase the accuracy of the determined concentration. The root mean squared errors of validation (RMSEV) of the validation process concerning alcohol and extract concentration were 0.23 Mas% (method A), 0.12 Mas% (method B), and 0.19 Mas% (method C) and 0.11 Mas% (method A), 0.11 Mas% (method B), and 0.11 Mas% (method C), respectively.

In an experimental study (October 2010 Mannheim Germany) with 99 Caucasian volunteers, the skin colour (L*, a*, b*) and the reflectance spectra of human skin were compared to the Fitzpatrick's sun-reactive skin photo types (SPT). For this purpose, the skin colour and the reflectance spectra of human skin were determined using non-invasive method with a newly developed fibre optic detection device. The device, based on reflectance spectroscopy, was designed and optimized using a commercial optical analysis Software. By means of the measured spectra of scattered light, the colour values and the absorption spectra of the skin were calculated. Neither any of the L*, a*, b* colour values nor any of the parameters of the absorbance spectra can be used alone to assess the skin type properly. Therefore, an ordinal logistic regression analysis was performed, using the statistical computing software r, to correlate the skin types with the measured optical parameters. It turned out that the detection device combined with the extended statistical analysis gives a better estimate of skin type in respect of the measured optical parameters than a procedure with only L*, a*, b* colour values. Even with the extended methodology, the procedure gives only a rough estimation of the skin type.