This is an R Markdown document. Markdown is a simple formatting syntax for authoring HTML, PDF, and MS Word documents. For more details on using R Markdown see http://rmarkdown.rstudio.com.
library(readxl)
infeccion <- read_excel("D:/Modelado/8 Oct/infeccion.xlsx",sheet = "afeccion")
infeccion
## # A tibble: 81 x 4
## enfermedad edadm pesoh17 genotipo
## <chr> <dbl> <dbl> <chr>
## 1 n 2 1 f
## 2 n 9 13 f
## 3 s 15 2 f
## 4 n 15 16 f
## 5 n 18 2 f
## 6 n 20 9 f
## 7 n 26 13 f
## 8 s 42 6 f
## 9 n 51 9 f
## 10 s 52 6 f
## # ... with 71 more rows
library(ggplot2)
par(mfrow=c(1,2))
boxplot(infeccion$pesoh17~infeccion$enfermedad,varwidth = TRUE, col=palette("Paired"))
boxplot(infeccion$edadm~infeccion$enfermedad,varwidth = TRUE, col=palette("Pastel 2"))
# S: si enferma
# N: no enferma
t<-table(infeccion$enfermedad,infeccion$genotipo)
addmargins(t)
##
## f m Sum
## n 17 47 64
## s 11 6 17
## Sum 28 53 81
plot(infeccion$edadm,infeccion$pesoh17, col="blue")
df=data.frame(infeccion)
df$enfermedad=ifelse(df$enfermedad=="n",0,1)
models<-glm(df$enfermedad~df$edadm*df$pesoh17*df$genotipo,family = "binomial")
summary(models)
##
## Call:
## glm(formula = df$enfermedad ~ df$edadm * df$pesoh17 * df$genotipo,
## family = "binomial")
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.1767 -0.5359 -0.2494 -0.1691 2.3149
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.109124 1.375388 -0.079 0.937
## df$edadm 0.024128 0.020874 1.156 0.248
## df$pesoh17 -0.074156 0.147678 -0.502 0.616
## df$genotipom -5.969109 4.278066 -1.395 0.163
## df$edadm:df$pesoh17 -0.001977 0.002006 -0.985 0.325
## df$edadm:df$genotipom 0.038086 0.041325 0.922 0.357
## df$pesoh17:df$genotipom 0.213830 0.343265 0.623 0.533
## df$edadm:df$pesoh17:df$genotipom -0.001651 0.003419 -0.483 0.629
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 83.234 on 80 degrees of freedom
## Residual deviance: 55.706 on 73 degrees of freedom
## AIC: 71.706
##
## Number of Fisher Scoring iterations: 6
modelr<-glm(df$enfermedad~1,family = "binomial")
summary(modelr)
##
## Call:
## glm(formula = df$enfermedad ~ 1, family = "binomial")
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.6864 -0.6864 -0.6864 -0.6864 1.7671
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.3257 0.2729 -4.859 1.18e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 83.234 on 80 degrees of freedom
## Residual deviance: 83.234 on 80 degrees of freedom
## AIC: 85.234
##
## Number of Fisher Scoring iterations: 4
anova(modelr,models,test="Chisq") # Se rechaza que el modelr es el mejor
## Analysis of Deviance Table
##
## Model 1: df$enfermedad ~ 1
## Model 2: df$enfermedad ~ df$edadm * df$pesoh17 * df$genotipo
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 80 83.234
## 2 73 55.706 7 27.529 0.0002676 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
models2<-glm(df$enfermedad~df$edadm*df$pesoh17+df$genotipo,family = "binomial")
summary(models2)
##
## Call:
## glm(formula = df$enfermedad ~ df$edadm * df$pesoh17 + df$genotipo,
## family = "binomial")
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.3085 -0.5477 -0.2679 -0.2170 2.2644
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.607709 1.249514 -0.486 0.6267
## df$edadm 0.029317 0.015014 1.953 0.0509 .
## df$pesoh17 -0.084943 0.124831 -0.680 0.4962
## df$genotipom -1.691206 0.714762 -2.366 0.0180 *
## df$edadm:df$pesoh17 -0.001719 0.001353 -1.271 0.2039
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 83.234 on 80 degrees of freedom
## Residual deviance: 58.142 on 76 degrees of freedom
## AIC: 68.142
##
## Number of Fisher Scoring iterations: 5
anova(models2,models,test="Chisq") # El mas peque昼㸱o es el mejor
## Analysis of Deviance Table
##
## Model 1: df$enfermedad ~ df$edadm * df$pesoh17 + df$genotipo
## Model 2: df$enfermedad ~ df$edadm * df$pesoh17 * df$genotipo
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 76 58.142
## 2 73 55.706 3 2.4368 0.4868
models3<-glm(df$enfermedad~df$edadm+df$pesoh17+df$genotipo,family = "binomial")
summary(models3)
##
## Call:
## glm(formula = df$enfermedad ~ df$edadm + df$pesoh17 + df$genotipo,
## family = "binomial")
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.9481 -0.5284 -0.3120 -0.1437 2.2525
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.609369 0.803288 0.759 0.448096
## df$edadm 0.012653 0.006772 1.868 0.061701 .
## df$pesoh17 -0.227912 0.068599 -3.322 0.000893 ***
## df$genotipom -1.543444 0.685681 -2.251 0.024388 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 83.234 on 80 degrees of freedom
## Residual deviance: 59.859 on 77 degrees of freedom
## AIC: 67.859
##
## Number of Fisher Scoring iterations: 5
anova(models2,models3,test="Chisq") # El mas peque昼㸱o es el mejor
## Analysis of Deviance Table
##
## Model 1: df$enfermedad ~ df$edadm * df$pesoh17 + df$genotipo
## Model 2: df$enfermedad ~ df$edadm + df$pesoh17 + df$genotipo
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 76 58.142
## 2 77 59.859 -1 -1.7161 0.1902
\[\ln(p/(1-p))=0.609+0.0126e-0.228w-1.543gen\] \[p/(1-p)=\exp(0.609+0.0126e-0.228w-1.543gen)\]
\[p=\exp(0.609+0.0126e-0.228w-1.543gen)(1-p)\] \[p+p(\exp(0.609+0.0126e-0.228w-1.543gen))=\exp(0.609+0.0126e-0.228w-1.543gen)\] \[p(1+(\exp(0.609+0.0126e-0.228w-1.543gen)))=\exp(0.609+0.0126e-0.228w-1.543gen)\]
\[\hat p= \frac{\exp(0.609+0.0126e-0.228w-1.543gen)}{1+(\exp(0.609+0.0126e-0.228w-1.543gen))}\]
df$genotipo=ifelse(df$genotipo=="f",0,1)
# f seria igual a 0
# m seria igual a 1
pestimado<-exp(0.609+0.0126*df$edadm-0.228*df$pesoh17-1.543*df$genotipo)/(1+exp(0.609+0.0126*df$edadm-0.228*df$pesoh17-1.543*df$genotipo))
pestimado
## [1] 0.600176361 0.096076499 0.584676265 0.054681317 0.593824864 0.233080134
## [7] 0.116355980 0.442801491 0.309939961 0.474073275 0.200431446 0.306996224
## [13] 0.129678994 0.785969591 0.138452844 0.115986349 0.175115156 0.272812422
## [19] 0.688260608 0.555125007 0.573708979 0.328053945 0.330837414 0.729561707
## [25] 0.466301165 0.849002647 0.356130232 0.462570155 0.012850992 0.010257862
## [31] 0.025150631 0.010257862 0.048642143 0.008286554 0.020376221 0.010648823
## [37] 0.099642780 0.009043718 0.014551540 0.047434909 0.038597710 0.013323669
## [43] 0.066248347 0.014899735 0.030560789 0.030936303 0.016051046 0.017269381
## [49] 0.402899089 0.235160095 0.464708781 0.029718975 0.031978206 0.078681336
## [55] 0.033575712 0.053860125 0.068319164 0.040279478 0.040769388 0.043305685
## [61] 0.043305685 0.084246515 0.476666960 0.133889229 0.112066939 0.386701189
## [67] 0.616802889 0.508149278 0.066583171 0.188069984 0.069784783 0.061966020
## [73] 0.064199914 0.127883871 0.105722348 0.130720869 0.051115466 0.130584570
## [79] 0.108012997 0.086827612 0.350145434
tablape<-ifelse(pestimado>0.5,1,0)
q<-table(df$enfermedad,tablape);q # Matriz de cofusion
## tablape
## 0 1
## 0 60 4
## 1 10 7
df$estimado=pestimado
tapply(df$estimado,df$genotipo,mean) # probabiidad estimada media x genotipo
## 0 1
## 0.3918100 0.1126742
tapply(df$enfermedad,df$genotipo,mean) # probabilidad enfermedad media x genotipo
## 0 1
## 0.3928571 0.1132075
plot(df$enfermedad,tablape,xlim=c(-0.1,1),ylim=c(-0.15,1),type='n', xlab = "",
ylab = "", yaxt='n', xaxt='n')
points(c(0,0,1,1),c(0,1,0,1),cex=c(60,4,10,7)*0.2,col=c("gold","blue","cyan","yellow"),pch=19)
Modelando proporciones
lote1=ifelse(round(rexp(n = 100,rate = 1.2))>=1,0,1)
prop.table(table(lote1))
## lote1
## 0 1
## 0.59 0.41
lote2=ifelse(round(rexp(n = 144,rate = 1.5))>=1,0,1)
prop.table(table(lote2))
## lote2
## 0 1
## 0.4930556 0.5069444
prevprome<-(0.36+0.486)/2 ; prevprome # erronea
## [1] 0.423
prevpromb<-(36+70)/(100+144) ; prevpromb # correcto
## [1] 0.4344262
machos<-c(0,1,3,4,33,80,158,365)
hembras<-c(1,3,7,18,22,41,52,79)
densidad<-c(1,4,10,22,55,121,210,444)
y<-cbind(machos,hembras);y
## machos hembras
## [1,] 0 1
## [2,] 1 3
## [3,] 3 7
## [4,] 4 18
## [5,] 33 22
## [6,] 80 41
## [7,] 158 52
## [8,] 365 79
summary(glm(y~densidad,family=binomial))
##
## Call:
## glm(formula = y ~ densidad, family = binomial)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -3.4619 -1.2760 -0.9911 0.5742 1.8795
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.0807368 0.1550376 0.521 0.603
## densidad 0.0035101 0.0005116 6.862 6.81e-12 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 71.159 on 7 degrees of freedom
## Residual deviance: 22.091 on 6 degrees of freedom
## AIC: 54.618
##
## Number of Fisher Scoring iterations: 4
summary(glm(y~log(densidad),family=binomial))
##
## Call:
## glm(formula = y ~ log(densidad), family = binomial)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.9697 -0.3411 0.1499 0.4019 1.0372
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.65927 0.48758 -5.454 4.92e-08 ***
## log(densidad) 0.69410 0.09056 7.665 1.80e-14 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 71.1593 on 7 degrees of freedom
## Residual deviance: 5.6739 on 6 degrees of freedom
## AIC: 38.201
##
## Number of Fisher Scoring iterations: 4
Modelo
\[\ln(p/(1-p))=-2.66+0.69\ln(density)\]
Estimando probabilidades
pobservado<-machos/(machos+hembras)
plot(log(densidad),pobservado,pch=16,col="green3")
pestimado<-exp(-2.66+0.69*log(densidad))/(1+exp(-2.66+0.69*log(densidad)))
pobservado/pestimado
## [1] 0.0000000 1.6232257 1.1756783 0.4898453 1.1401624 1.0066236 1.0211194
## [8] 0.9972337
plot(pobservado,pestimado,col="blue3")
plot(log(densidad),pobservado,pch=16,col="blue")
points(log(densidad),pestimado,pch=16,col="cyan")
segments(x0 =log(densidad)[4] ,y0 =pobservado[4] ,x1 =log(densidad)[4] ,y1 =pestimado[4])
text(3.8,0.25, "M攼㸱ximo \n residual")
res<-pobservado-pestimado
boxplot(res, col = "cyan")
outliers::dixon.test(res)
##
## Dixon test for outliers
##
## data: res
## Q = 0.4712, p-value = 0.2132
## alternative hypothesis: lowest value -0.189356501197036 is an outlier
Hipotesis nula: el valor mas bajo no es un atipico Hipotesis alterna: el valor mas bajo es un atipico Por lo tanto no se rechaza la hipotesis nula (No es un atipico)
library(mvoutlier)
## Loading required package: sgeostat
## Registered S3 method overwritten by 'GGally':
## method from
## +.gg ggplot2
## sROC 0.1-2 loaded
z <- cbind(log(densidad),pobservado)
# execute:
color.plot(z, quan=0.95)
## $outliers
## [1] FALSE FALSE FALSE TRUE FALSE FALSE FALSE FALSE
##
## $md
## [1] 1.3644872 1.3103553 0.9787861 6.9571001 1.0225690 0.5371411 0.7275172
## [8] 1.1303935
##
## $euclidean
## [1] 0.1420777 1.1793198 1.5883896 1.6629228 2.8765799 3.2845987 3.6866598
## [8] 4.1048014
corr.plot(z[,1],z[,2])
## $cor.cla
## [1] 0.9488771
##
## $cor.rob
## [1] 0.9925565
library(readxl)
germination <- read_excel("D:/Modelado/8 Oct/germination.xlsx", sheet = "germinacion")
germination
## # A tibble: 21 x 4
## nogerm sigerm tiempo extracto
## <dbl> <dbl> <dbl> <dbl>
## 1 10 39 40 1
## 2 23 62 40 1
## 3 23 81 40 1
## 4 26 51 40 1
## 5 17 39 40 1
## 6 5 6 40 2
## 7 53 74 40 2
## 8 55 72 40 2
## 9 32 51 40 2
## 10 46 79 40 2
## # ... with 11 more rows
germination$pgermi<-germination$sigerm/(germination$nogerm+germination$sigerm)
boxplot(germination$pgermi, col = "green")
germination$germi1<-sqrt(germination$pgermi)
germination$germi2<-asin(germination$germi1)
m1<-aov(germination$pgermi~germination$extracto*germination$tiempo)
summary(m1)
## Df Sum Sq Mean Sq F value Pr(>F)
## germination$extracto 1 0.07685 0.07685 11.518 0.00345 **
## germination$tiempo 1 0.02685 0.02685 4.024 0.06105 .
## germination$extracto:germination$tiempo 1 0.00522 0.00522 0.783 0.38868
## Residuals 17 0.11343 0.00667
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
germination$trt=interaction(germination$tiempo,germination$extracto)
bartlett.test(germination$pgermi,germination$trt)
##
## Bartlett test of homogeneity of variances
##
## data: germination$pgermi and germination$trt
## Bartlett's K-squared = 8.374, df = 3, p-value = 0.03888
shapiro.test(m1$residuals)
##
## Shapiro-Wilk normality test
##
## data: m1$residuals
## W = 0.90092, p-value = 0.03646
m2<-aov(germination$germi1~germination$extracto*germination$tiempo)
summary(m2)
## Df Sum Sq Mean Sq F value Pr(>F)
## germination$extracto 1 0.02749 0.027487 12.802 0.00232 **
## germination$tiempo 1 0.00954 0.009538 4.442 0.05021 .
## germination$extracto:germination$tiempo 1 0.00244 0.002436 1.135 0.30169
## Residuals 17 0.03650 0.002147
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
bartlett.test(germination$germi1,germination$trt)
##
## Bartlett test of homogeneity of variances
##
## data: germination$germi1 and germination$trt
## Bartlett's K-squared = 6.7777, df = 3, p-value = 0.07933
shapiro.test(m2$residuals)
##
## Shapiro-Wilk normality test
##
## data: m2$residuals
## W = 0.92054, p-value = 0.08897
# Usando la segunda transformaci昼㸳n
m3<-aov(germination$germi2~germination$extracto*germination$tiempo)
summary(m3)
## Df Sum Sq Mean Sq F value Pr(>F)
## germination$extracto 1 0.12622 0.12622 6.867 0.0179 *
## germination$tiempo 1 0.05257 0.05257 2.860 0.1090
## germination$extracto:germination$tiempo 1 0.00036 0.00036 0.019 0.8910
## Residuals 17 0.31248 0.01838
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
bartlett.test(germination$germi2,germination$trt)
##
## Bartlett test of homogeneity of variances
##
## data: germination$germi2 and germination$trt
## Bartlett's K-squared = 17.59, df = 3, p-value = 0.0005343
shapiro.test(m3$residuals)
##
## Shapiro-Wilk normality test
##
## data: m3$residuals
## W = 0.77241, p-value = 0.0002558
AIC(m1);AIC(m2);AIC(m3)
## [1] -40.04684
## [1] -63.85815
## [1] -18.76702
pred2<-m2$fitted.values
plot(germination$pgermi,pred2, col = "red")
Usando Modelo Lineal Generalizado GLM
y2<-cbind(germination$nogerm,germination$sigerm)
summary(glm(y2~germination$tiempo*germination$extracto,family = binomial))
##
## Call:
## glm(formula = y2 ~ germination$tiempo * germination$extracto,
## family = binomial)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.63783 -0.85274 0.03306 0.51772 1.37050
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.78122 0.77575 -1.007 0.314
## germination$tiempo -0.02141 0.02225 -0.962 0.336
## germination$extracto -0.04883 0.46690 -0.105 0.917
## germination$tiempo:germination$extracto 0.01690 0.01333 1.267 0.205
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 33.870 on 20 degrees of freedom
## Residual deviance: 13.079 on 17 degrees of freedom
## AIC: 105.05
##
## Number of Fisher Scoring iterations: 4
mo1<-glm(y2~germination$tiempo*germination$extracto,family = quasibinomial)
summary(mo1)
##
## Call:
## glm(formula = y2 ~ germination$tiempo * germination$extracto,
## family = quasibinomial)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.63783 -0.85274 0.03306 0.51772 1.37050
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.78122 0.64788 -1.206 0.244
## germination$tiempo -0.02141 0.01858 -1.152 0.265
## germination$extracto -0.04883 0.38994 -0.125 0.902
## germination$tiempo:germination$extracto 0.01690 0.01114 1.517 0.148
##
## (Dispersion parameter for quasibinomial family taken to be 0.697491)
##
## Null deviance: 33.870 on 20 degrees of freedom
## Residual deviance: 13.079 on 17 degrees of freedom
## AIC: NA
##
## Number of Fisher Scoring iterations: 4
mo2<-glm(y2~germination$tiempo+germination$extracto,family = quasibinomial)
summary(mo2)
##
## Call:
## glm(formula = y2 ~ germination$tiempo + germination$extracto,
## family = quasibinomial)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.5431 -0.5006 -0.1852 0.3968 1.4796
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -1.695438 0.263091 -6.444 4.6e-06 ***
## germination$tiempo 0.005641 0.005687 0.992 0.334391
## germination$extracto 0.523241 0.106982 4.891 0.000118 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasibinomial family taken to be 0.7522366)
##
## Null deviance: 33.870 on 20 degrees of freedom
## Residual deviance: 14.678 on 18 degrees of freedom
## AIC: NA
##
## Number of Fisher Scoring iterations: 4
anova(mo1,mo2,test="F")
## Analysis of Deviance Table
##
## Model 1: y2 ~ germination$tiempo * germination$extracto
## Model 2: y2 ~ germination$tiempo + germination$extracto
## Resid. Df Resid. Dev Df Deviance F Pr(>F)
## 1 17 13.079
## 2 18 14.678 -1 -1.5995 2.2933 0.1483
mo3<-glm(y2~germination$extracto,family = quasibinomial)
summary(mo3)
##
## Call:
## glm(formula = y2 ~ germination$extracto, family = quasibinomial)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.59555 -0.69284 -0.04085 0.51537 1.25485
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -1.5061 0.1791 -8.408 7.95e-08 ***
## germination$extracto 0.5244 0.1064 4.927 9.36e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasibinomial family taken to be 0.7450712)
##
## Null deviance: 33.870 on 20 degrees of freedom
## Residual deviance: 15.422 on 19 degrees of freedom
## AIC: NA
##
## Number of Fisher Scoring iterations: 4
anova(mo3,mo2,test="F")
## Analysis of Deviance Table
##
## Model 1: y2 ~ germination$extracto
## Model 2: y2 ~ germination$tiempo + germination$extracto
## Resid. Df Resid. Dev Df Deviance F Pr(>F)
## 1 19 15.422
## 2 18 14.678 1 0.74394 0.989 0.3332
mo3$fitted.values
## 1 2 3 4 5 6 7 8
## 0.2725599 0.2725599 0.2725599 0.2725599 0.2725599 0.3876404 0.3876404 0.3876404
## 9 10 11 12 13 14 15 16
## 0.3876404 0.3876404 0.3876404 0.2725599 0.2725599 0.2725599 0.2725599 0.2725599
## 17 18 19 20 21
## 0.3876404 0.3876404 0.3876404 0.3876404 0.3876404
plot((germination$nogerm/(germination$nogerm+germination$sigerm)),mo3$fitted.values, col = "blue2")
plot((germination$sigerm/(germination$nogerm+germination$sigerm)),mo3$fitted.values, col = "green4")