What are the Inputs Factors influencing ANOVA (GLM) Results in an Production Optimization experiment?
I want to predict the important production parameters affecting my production process (Sheet metal Rolling) and product end properties (Yield Strength - Dependent Parameter). The independent parameters are, Chemical Comp. - Si, Mg, Fe etc, Hot Rolling Temp, Homogenization Temp. etc. All my indepent parameters have different levels (Some times, some factors might have up to 35 Levels)
I used ANOVA-GLM in minitab for my analysis. The data I used for the analysis belongs to the last two years production data.
1.) The problem is, when I use the complete two years process data for my analysis, I get different % contribution in comparison to results for which I used only a part of the two years data. With smaller no. of datas I get a better model which explains 95% of the variance. Whereas the other with large set of datas explains only 65% of varaince. So, I want to know which one to accept / beleive and what is the recommended no. of data to use that would replicate the behaviour of my process?
2.) What is the effect of using, duplicate datas (Same data repeated twice or thrice or more) in the analysis.?
3. ) For all of my analysis i have got PRESS, RSq(pred) as *. What does this mean.?
4. ) Are there any other tool / method that would be suitable for my data.? I think taguchi cannot be used, due to its constrains with usage levels.
22 фрілансерів(-и) у середньому готові виконати цю роботу за $23/годину
I am a statistical analyst, with strong background of Pure Statistics including a strong grip on statistical software including SPSS, Excel, Minitab, STATA, SAS, Mathematica, Statistica and Eviews.
Hello I am expert Statistics instructor and consultant. I am expert in Minitab 14, 15, 16, 17 & 18. I am professional in Design of Experiment and ANOVA.