Coating - learn how to calculate and analyse a PLS model

Task To create one or several models describing the variation in the dependent variables BRIGHT, OPAC and GLOSS as good as possible. Investigate if there are groups in the material and what causes them. Data The data set contains material from an incomplete, designed experiment. The experiment was designed to study ideal levels for five independent variables (STARCH, LATEX, HP91, COATWT and CPSI) that are necessary to manage the minimum levels for the dependent variables (BRIGHT, OPAC, GLOSS). In the data set STARCH, LATEX, HP91 and COATWT are varied on five levels and CPSI is varied on two levels. The independent variables could be adjusted and maintained with the desired value for STARCH, LATEX, HP91 and CPSI while COATWT couldn’t be controlled with the same accuracy. The set point for COATWT has been replaced with measured values. You will find the data in COATING.XLS.

1. Open COATING.XLS 2. Draw variables 3. Calculate the model 4. Analyse the loadingplot, part 1 5. Analyse the loadingplot, part 2 6. Analyse the scoreplot 7. Draw prediction error 8. Draw contributions 9. Draw predictions 10. Draw residuals 11. Draw regression coefficients 12. Analyse the correlation between the Y variables 13. Save the document