John Fox
2005-Apr-18 23:08 UTC
[R] Discrepancy between gam from gam package and gam in S-PLUS
Dear Trevor, I've noticed a discrepancy in the degrees of freedom reported by gam() from the gam package in R vs. gam() in S-PLUS. The nonparametric df differ by 1; otherwise (except for things that depend upon the df), the output is the same: --------- snip ------------ *** From R (gam version 0.93):> mod.gam <- gam(prestige ~ lo(income, span=.6), data=Prestige) > summary(mod.gam)Call: gam(formula = prestige ~ lo(income, span = 0.6), data = Prestige) Deviance Residuals: Min 1Q Median 3Q Max -17.163 -8.205 -1.998 8.070 32.326 (Dispersion Parameter for gaussian family taken to be 122.5047) Null Deviance: 29895.43 on 101 degrees of freedom Residual Deviance: 12004.72 on 97.9939 degrees of freedom AIC: 785.82 Number of Local Scoring Iterations: 2 DF for Terms and F-values for Nonparametric Effects Df Npar Df Npar F Pr(F) (Intercept) 1 lo(income, span = 0.6) 1 2 10.626 6.537e-05 *** --- Signif. codes: 0 `***' 0.001 `**' 0.01 `*' 0.05 `.' 0.1 ` ' 1 *** From S-PLUS 6.2.1 for Windows:> library(car) > mod.gam <- gam(prestige ~ lo(income, span=.6), data=Prestige) > summary(mod.gam)Call: gam(formula = prestige ~ lo(income, span = 0.6), data = Prestige) Deviance Residuals: Min 1Q Median 3Q Max -17.16264 -8.205246 -1.997525 8.070375 32.32608 (Dispersion Parameter for Gaussian family taken to be 123.7677 ) Null Deviance: 29895.43 on 101 degrees of freedom Residual Deviance: 12004.72 on 96.99392 degrees of freedom Number of Local Scoring Iterations: 1 DF for Terms and F-values for Nonparametric Effects Df Npar Df Npar F Pr(F) (Intercept) 1 lo(income, span = 0.6) 1 3 7.018995 0.0002517358>--------- snip ------------ I suppose that one of these must be in error. Regards, John -------------------------------- John Fox Department of Sociology McMaster University Hamilton, Ontario Canada L8S 4M4 905-525-9140x23604 http://socserv.mcmaster.ca/jfox