This package compute confidence intervals on individual variances for any balanced mixed model, involving five or fewer factors according to Burdick and Graybill (1992) methodology. The package would fit a saturated model to the data (main effect and interaction terms of all orders) providing an ANOVA model with Expected Mean Squares (type III) by using the (Bennett and Franklin 1954) algorithm. The output will include: a table showing the confidence interval (CI) and estimate of each covariance parameter, as well as the method used to compute the CI, the Expected Mean Square, an ANOVA table showing the degrees of freedom (df), sums of squares (SS), mean squares (MS), F statistic, and finally AIC and BIC measures.
| Version: | 1.0.1 |
| Depends: | methods |
| Published: | 2012-05-05 |
| Author: | Civit, S., Vilardell, M., Hess, A., Matthew, Z., Ge, Y., Caballe, A. |
| Maintainer: | Sergi Civit <svives at ub.edu> |
| License: | GPL (≥ 2) |
| CRAN checks: | varcompci results |
| Package source: | varcompci_1.0.1.tar.gz |
| MacOS X binary: | varcompci_1.0.1.tgz |
| Windows binary: | varcompci_1.0.1.zip |
| Reference manual: | varcompci.pdf |
| Vignettes: |
Gene Response Characterization of AKT inhibitors Batch Process |
| Old sources: | varcompci archive |