Identifying cause of quality defect in cheese using qualitative variables in a statistical experimental design
2004 (English)In: Journal of Chemometrics, ISSN 0886-9383, Vol. 18, no 3-4, 139-45 p.Article in journal (Refereed) Published
Statistical experimental design was used to identify the stage or stages in a cheesemaking process that caused a quality defect. A full factorial design with five qualitative variables was constructed and a total of 32 experiments, each representing a unique combination of treatments, were then evaluated by a trained sensory panel. The average score for each sample was used as a response, and classical multiple linear regression (MLR) was used to find the relationship between the treatments (X) and the rate of quality defect (y). Evaluation of the design and the regression coefficients clearly showed that one of the factors, process step 3 (x3) at location B, was mainly responsible for high rates of the quality defect. In addition, process step 5 at location B, especially in combination with process step 2 or 4 at location A or step 3 at location B, increased the rate of quality defect in the cheese. The other factors had little or no effect.
Place, publisher, year, edition, pages
2004. Vol. 18, no 3-4, 139-45 p.
statistical experimental design, MLR, sensory evaluation, quality defect, product and process optimization
IdentifiersURN: urn:nbn:se:umu:diva-14194DOI: 10.1002/cem.844OAI: oai:DiVA.org:umu-14194DiVA: diva2:153865