Hi, I am running a simulation and have to perform ANOVA to determine the rank of factors. Used the aov() function and it works great for full factorial design.
1. For a massive set of data, I tried using biglm, while it can create the linear model, all the residuals (for assumption validation) are not recorded and the sum of squares are not there, just the estimated regression coefficient, 95% CI, SE and p. Can I use any of these to get the rank of factors ? 2. I'm trying to use Latin Hypercube design instead of the costly full factorial design. However, if I choose 2 partitions with 2 variables (for experiment with 2 factors - A & B each with 2 levels - min & max), I could not use aov() to get the rank of factors since aov() detects that B is "dependant" to A, thus only A causes the variance. e.g: Design point 1: A (min), B (max) Design point 2: A (max), B(min) Terms: Sum of Squares 34.83342 (A) 0.96427 (Residuals) Deg. of Freedom 1 (A) 198 (Residuals) Residual standard error: 0.06978563 2 out of 4 effects are not estimable Estimated effects may be unbalanced Please advice how to solve this problem. Thank you, Hardi ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
