logreg

setwd("D:/logreg/")
data <- read.table("data.txt", sep = ",", header = TRUE)
data <- data[, 1:5]
# Задаем факторы.
data[, c(1, 3:5)] <- lapply(data[, c(1, 3:5)], as.factor)

Предиктор - методика

fit <- glm(группа ~ методика, family = binomial(link = "logit"), data = data)
summary(fit)

## 
## Call:
## glm(formula = группа ~ методика, family = binomial(link = "logit"), 
##     data = data)
## 
## Deviance Residuals: 
##      Min        1Q    Median        3Q       Max  
## -1.99783  -0.61352   0.02096   0.83613   1.43882  
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)    
## (Intercept) -2.55301    0.45079  -5.663 1.48e-08 ***
## методика     0.48918    0.07698   6.355 2.09e-10 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 249.53  on 179  degrees of freedom
## Residual deviance: 193.53  on 178  degrees of freedom
## AIC: 197.53
## 
## Number of Fisher Scoring iterations: 4

library(ROCR)

## Loading required package: gplots

## 
## Attaching package: 'gplots'

## The following object is masked from 'package:stats':
## 
##     lowess

pred <- predict(fit, newdata = data, type = "response")
pr <- prediction(pred, data$группа)
prf <- performance(pr, measure = "tpr", x.measure = "fpr")
plot(prf)

auc <- performance(pr, measure = "auc")
auc <- auc@y.values[[1]]
auc

## [1] 0.7680864

Предикторы - методика + три переменные

fit <- glm(группа ~ ., family = binomial(link = "logit"), data = data)
summary(fit)

## 
## Call:
## glm(formula = группа ~ ., family = binomial(link = "logit"), 
##     data = data)
## 
## Deviance Residuals: 
##      Min        1Q    Median        3Q       Max  
## -2.76676  -0.48502  -0.04623   0.74480   1.81180  
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)    
## (Intercept) -3.06332    0.55664  -5.503 3.73e-08 ***
## методика     0.32747    0.09269   3.533 0.000411 ***
## дихотом11    0.82448    0.41452   1.989 0.046703 *  
## дихотом21    1.54216    0.41898   3.681 0.000233 ***
## порядковая1  0.36942    0.52253   0.707 0.479576    
## порядковая2  1.88240    0.61178   3.077 0.002092 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 249.53  on 179  degrees of freedom
## Residual deviance: 160.41  on 174  degrees of freedom
## AIC: 172.41
## 
## Number of Fisher Scoring iterations: 5

library(ROCR)
pred <- predict(fit, newdata = data, type = "response")
pr <- prediction(pred, data$группа)
prf <- performance(pr, measure = "tpr", x.measure = "fpr")
plot(prf)

auc <- performance(pr, measure = "auc")
auc <- auc@y.values[[1]]
auc

## [1] 0.8666667