R glm

R 逻辑回归

[zhaosl@Linuxmg 20180354]$ cat data
gc  1/1 1/2 2/2
case    356 443 136
contrl  158 128 37

R 怎么做逻辑回归

a = read.table("data", header= T, sep = "\t", stringsAsFactors = F)
tt1 <- glm(as.matrix(a[2:3])~c(1,0), family=binomial(link=logit))

> summary(tt1)

Call:
glm(formula = as.matrix(a[2:3]) ~ c(1, 0), family = binomial(link = logit))

Deviance Residuals: 
[1]  0  0

Coefficients:
            Estimate Std. Error z value Pr(>|z|)   
(Intercept)   0.2106     0.1189   1.771  0.07662 . 
c(1, 0)      -0.4292     0.1386  -3.097  0.00196 **
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance:  9.6487e+00  on 1  degrees of freedom
Residual deviance: -1.5810e-13  on 0  degrees of freedom
AIC: 17.222

Number of Fisher Scoring iterations: 2

#P值
> summary(tt1)$coefficients
              Estimate Std. Error   z value    Pr(>|z|)
(Intercept)  0.2105648  0.1189185  1.770664 0.076616549
c(1, 0)     -0.4292038  0.1385927 -3.096871 0.001955746

> summary(tt1)$coefficients[2,4] == summary(tt1)$coefficients[8]
[1] TRUE

#OR值
> exp(coef(tt1))
(Intercept)     c(1, 0) 
  1.2343750   0.6510272

> exp(coef(tt1))[2]
  c(1, 0) 
0.6510272

#置信区间CI
exp(confint(tt1))
Waiting for profiling to be done...
                2.5 %    97.5 %
(Intercept) 0.9784116 1.5602708
c(1, 0)     0.4956864 0.8536914

> exp(confint(tt1))[2,]
Waiting for profiling to be done...
    2.5 %    97.5 % 
0.4956864 0.8536914