The application of multivariate mathematical-statistical methods to the solution of problems in nutrition theory, nutritional economics and food technology. Part 5. Cluster analysis as a tool for proofing sensory attributes

Autor: M. Rothe, V. Erhardt, M. Specht, B. Seppelt
Rok vydání: 1986
Předmět:
Zdroj: Food / Nahrung. 30:119-130
ISSN: 1521-3803
0027-769X
DOI: 10.1002/food.19860300203
Popis: The task of the test program was the replacement of more subjective and time-consuming sensory methods by instrumental or chemical ones. In the case of meat flavour concentrates (MFC) and their burnt-bitter flavour the cluster analysis resulted in the following: –The test program was based on 20 samples produced by statistical plans (central composite design for 3 explanatory variables). Starting with 10 sensory and 40 instrumental methods the latter have been reduced to 10 methods considering the suppositions for application of the cluster analysis. –The 20 samples were tested on one side by a trained sensory panel for the intensity of the “burnt-bitter” flavour and parallely by the selected 10 instrumental values. This resulted in both cases in the differentiation into 3 clusters with significant differences. –For evaluation of MFC as to their “burnt-bitter” flavour it was shown that the sensory evaluation can be replaced or supported by analytical data successfully. Thus cluster analysis is a useful tool for this purpose delivering transparent results.
Databáze: OpenAIRE
Externí odkaz: