PCA application - PowerPoint PPT Presentation

1 / 8
About This Presentation
Title:

PCA application

Description:

Food Quality and Pref. 1, 29-36. Practical problem. Processed cheese ... Can we increase meltability and decrease stickiness, graininess and adhesiveness ... – PowerPoint PPT presentation

Number of Views:22
Avg rating:3.0/5.0
Slides: 9
Provided by: torm8
Category:
Tags: pca | application | pref

less

Transcript and Presenter's Notes

Title: PCA application


1
PCA application
  • Processed cheese
  • Ref Ellekjær, Ilseng and Næs (1996). A case
    study of the use of exp. design and mult. anal.
    in product development. Food Quality and Pref. 1,
    29-36.

2
Practical problem
  • Processed cheese
  • Improvement of texture properties
  • Can we increase meltability and decrease
    stickiness, graininess and adhesiveness

3
Designed experiment
  • 7 factors
  • Melting salt, maturity of cheese, dry matter of
    processed cheese
  • pH, addition of dry matter, after creaming,
    cooling of cheese
  • Both process and ingredient variables
  • Fractional factorial design, 27-2 32
  • Three center points for each melting salt
  • Three replicates of a reference were added
    (original recipee)
  • Resolution IV design
  • Main effects and two factors interactions are not
    confounded
  • Sensory responses
  • Glossiness, ability to retain shape,
    adhesiveness, firmness, graininess, stickiness,
    meltiness, condensation

4
Analysis techniques
  • Used averages of sensory variables (over
    assessors)
  • Used PCA and ANOVA
  • PCA gives an overview and indicates optimal
    samples in the design
  • ANOVA is important for determining which
    variables that are important for further study

5
ANOVA results
  • Higher order interactions used to test
    significance
  • Most important variables were maturity of
    cheese, dry matter of processed cheese and pH
  • Only a few interactions were significant

6
Two first PCs
Best candidates are marked with green
Positive attribute green Negative attribute
-red
7
PC components 2 and 3
Best candidates are marked with green
Positive attribute green Negative attribute
-red
8
Conclusions
  • Replicates quite close
  • Supports reliability of experiment
  • We got information about which samples that are
    best
  • We got information about which experimental
    factors that are most important
Write a Comment
User Comments (0)
About PowerShow.com