> foedselsvekt <- read.table("foedselsvekt",col.names=c("kjonn","varighet","vekt")) > foedselsvekt kjonn varighet vekt 1 1 40 2968 2 1 38 2795 3 1 40 3163 4 1 35 2925 5 1 36 2625 6 1 37 2847 7 1 41 3292 8 1 40 3473 9 1 37 2628 10 1 38 3176 11 1 40 3421 12 1 38 2975 13 2 40 3317 14 2 36 2729 15 2 40 2935 16 2 38 2754 17 2 42 3210 18 2 39 2817 19 2 40 3126 20 2 37 2539 21 2 36 2412 22 2 38 2991 23 2 39 2875 24 2 40 3231 > names(foedselsvekt) [1] "kjonn" "varighet" "vekt" > foedselsvekt$kjonn [1] 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 > foedselsvekt$kjonn <- factor(foedselsvekt$kjonn) > foedsel.res <- glm(vekt~kjonn+kjonn:varighet-1,family=gaussian(),data=foedselsvekt) > par(mfrow=c(2,1)) > class(foedsel.res) [1] "glm" "lm" > plot(foedsel.res) > foedsel.res2 <- glm(vekt~kjonn+varighet-1,family=gaussian(),data=foedselsvekt) > anova(foedsel.res,foedsel.res2,test="F") Analysis of Deviance Table Response: vekt Terms Resid. Df Resid. Dev Test Df Deviance F Value Pr(F) 1 kjonn + kjonn:varighet - 1 20 652424.5 2 kjonn + varighet - 1 21 658770.7 1 vs. 2 -1 -6346.225 0.1945428 0.6638934 > summary.aov(foedsel.res) Df Sum of Sq Mean Sq F Value Pr(F) kjonn 2 211445253 105722627 3240.915 0.000000e+00 kjonn:varighet 2 1101286 550643 16.880 5.078358e-05 Residuals 20 652425 32621 > biller <- read.table("biller",col.names=c("dose","antall","dode")) > biller dose antall dode 1 1.6907 59 6 2 1.7242 60 13 3 1.7552 62 18 4 1.7842 56 28 5 1.8113 63 52 6 1.8369 59 53 7 1.8610 62 61 8 1.8839 60 60 > biller.res <- glm(cbind(dode,antall-dode)~dose,family=binomial(link=logit), + data=biller) > summary(biller.res) Call: glm(formula = cbind(dode, antall - dode) ~ dose, family = binomial(link = logit ), data = biller) Deviance Residuals: Min 1Q Median 3Q Max -1.594122 -0.3943905 0.8329144 1.259226 1.593991 Coefficients: Value Std. Error t value (Intercept) -60.71733 5.173518 -11.73618 dose 34.27026 2.908076 11.78451 (Dispersion Parameter for Binomial family taken to be 1 ) Null Deviance: 284.2024 on 7 degrees of freedom Residual Deviance: 11.23223 on 6 degrees of freedom Number of Fisher Scoring Iterations: 3 Correlation of Coefficients: (Intercept) dose -0.9996809 > names(biller.res) [1] "coefficients" "residuals" "fitted.values" [4] "effects" "R" "rank" [7] "assign" "df.residual" "weights" [10] "prior.weights" "family" "linear.predictors" [13] "deviance" "null.deviance" "call" [16] "iter" "y" "contrasts" [19] "terms" "formula" > coef(biller.res) (Intercept) dose -60.71733 34.27026 > biller.res$coef (Intercept) dose -60.71733 34.27026 > residuals(biller.res) 1 2 3 4 5 6 7 8 1.283666 1.05968 -1.196117 -1.594122 0.6061488 -0.1271483 1.25108 1.593991 > par(mfrow=c(2,1)) > plot(biller.res) > predict(biller.res,biller) 1 2 3 4 5 6 7 8 -2.776609 -1.628556 -0.5661776 0.4276598 1.356384 2.233702 3.059616 3.844404 > > predict(biller.res,biller,type="response") 1 2 3 4 5 6 7 8 0.05860134 0.1640283 0.3621193 0.6053147 0.7951713 0.9032354 0.9551958 0.9790492 > lungekreft <- read.table("lungekreft",col.names=c("by","alder","kreft","befolkning")) > attach(lungekreft,1) > search() [1] "lungekreft" [2] "/mn/gilgamesj/u5/geirs/.Data" [3] "/local/src/SPlus/library/local/.Data" [4] "/local/src/SPlus/splus/.Functions" [5] "/local/src/SPlus/stat/.Functions" [6] "/local/src/SPlus/s/.Functions" [7] "/local/src/SPlus/s/.Datasets" [8] "/local/src/SPlus/stat/.Datasets" [9] "/local/src/SPlus/splus/.Datasets" [10] "/local/src/SPlus/library/boot/.Data" > objects() [1] ".Last.value" "alder" "befolkning" "by" "kreft" > by <- factor(by);alder <- factor(alder) > options(contrasts="contr.treatment") > logbef <- log(befolkning) > kreft.res1 <- glm(kreft~offset(logbef)+alder+by-1,family=poisson(link=log)) > kreft.res1 Call: glm(formula = kreft ~ offset(logbef) + alder + by - 1, family = poisson(link = log)) Coefficients: alder1 alder2 alder3 alder4 alder5 alder6 by2 -5.632065 -4.531051 -4.113452 -3.864358 -3.775201 -4.212411 -0.33006 by3 by4 -0.3715462 -0.2723177 Degrees of Freedom: 24 Total; 15 Residual Residual Deviance: 23.44748 > anova(kreft.res1,test="Chi") Analysis of Deviance Table Poisson model Response: kreft Terms added sequentially (first to last) Df Deviance Resid. Df Resid. Dev Pr(Chi) NULL 24 628.3534 alder 6 600.0469 18 28.3065 0.000000 by 3 4.8590 15 23.4475 0.182414