로지스틱 회귀모형

Code(1)

Sys.setenv(JAVA_HOME='C:\\Program Files\\Java\\Jdk1.7.0_79')
library(xlsx)
mower.data <- read.xlsx("c:/ian_R/mower.xlsx",1)
head(mower.data)
mower.logit <- glm(owner~.,family=binomial, data=mower.data)
summary(mower.logit)
 

1-pchisq(15.323, 21)

 

Result(1)

Call:
glm(formula = owner ~ ., family = binomial, data = mower.data)

Deviance Residuals: 
     Min        1Q    Median        3Q       Max  
-1.74044  -0.29685   0.00439   0.44750   1.86821  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)  
(Intercept) -25.9382    11.4871  -2.258   0.0239 *
income        0.3326     0.1629   2.042   0.0412 *
size          1.9276     0.9256   2.083   0.0373 *
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 33.271  on 23  degrees of freedom

    상수항만이 포함된 모형으로 적합할 대 모형추정 값과 관찰값의 차이에 관한 통계량

    Residual deviance: 15.323 on 21 degrees of freedom

    독립변수가 포함될 때의 차이에 대한 통계량 AIC: 21.323 Number of Fisher Scoring iterations: 6

 

 

 

1-pchisq(15.323, 21)

결과에서 residual deviance = 15.323이고, 자유도 21에서 유의확률 p-값 = 0.806

따라서 귀무가설 적합

 

 

 

 

 

 

 

Code(2) : 새로운 자료 분류

 

mower.predict <-predict(mower.logit, newdata = mower.data, type="response")
pred <- ifelse(mower.predict < 0.5, "no", "yes")
pred <- factor(pred)
confusion.matrix <- table(mower.data$owner, pred)
error <- 1-(sum(diag(confusion.matrix))/sum(confusion.matrix))
error

 

 

Result(2)

 

0.1666667

 

 

 

 

 

Code(1)Sys.setenv(JAVA_HOME='C:\\Program Files\\Java\\Jdk1.7.0_79')
library(xlsx)

 

 

Result(2)