#Đưa dữ liệu vào R
library(haven)
e1 <- read_sav("D:/Deskop/Endobronchial_biopsy/Xử lý số liệu/RStudio/Dữ liệu-FINAL.sav")
head(e1)
## # A tibble: 6 × 108
##   sothutu hoten   namsinh  gioi  ldvv V6    hutthuocla goinam   tha suytim   dtd
##     <dbl> <chr>     <dbl> <dbl> <dbl> <chr> <dbl+lbl>   <dbl> <dbl>  <dbl> <dbl>
## 1       1 LY VAN…    1956     1     5 ""    1 [Còn hú…     20     1      0     0
## 2       2 PHAM V…    1956     1     1 ""    0 [Không …     NA     1      0     0
## 3       3 HO VAN…    1957     1     6 "HO … 0 [Không …     NA     0      0     0
## 4       4 KIM CH…    1970     1     1 ""    1 [Còn hú…     37     0      0     0
## 5       5 TRAN T…    1969     1     6 "CO … 1 [Còn hú…     40     0      0     0
## 6       6 DANH T…    1976     0     1 ""    0 [Không …     NA     0      0     0
## # ℹ 97 more variables: ckd <dbl>, xogan <dbl>, doiquy <dbl>, ungthu <dbl>,
## #   loaiut <chr>, copd <dbl>, hen <dbl>, lao <dbl>, laomp <dbl>, cc <dbl>,
## #   cn <dbl>, sot <dbl>, met <dbl>, sutcan <dbl>, chanan <dbl>, hokhan <dbl>,
## #   hodam <dbl>, homau <dbl>, daunguc <dbl>, khotho <dbl>, khangiong <dbl>,
## #   nuotnghen <dbl>, sohach <dbl>, hc3giam <dbl>, tn <dbl>, tt <dbl>, nt <dbl>,
## #   khonggn <dbl>, kinhmo <dbl>, hang <dbl>, notmo <dbl>, dam <dbl>,
## #   dongdac <dbl>, xep <dbl>, tdmp <dbl>, trenp <dbl>, giuap <dbl>, …
e1$gpb1 <- factor(e1$gpb,
                   levels = c(1, 0),  # chú ý: 1 là UTP, 0 là không UTP
                   labels = c("UTP", "Không UTP"))
e1$nhomtuoi1 <- factor(e1$nhomtuoi, levels = c(1, 2, 3), labels = c("<50", "50-70", ">70"))

model_nhomtuoi1.1 <- glm(gpb ~ nhomtuoi1, data = e1, family = binomial)
summary(model_nhomtuoi1.1)
## 
## Call:
## glm(formula = gpb ~ nhomtuoi1, family = binomial, data = e1)
## 
## Coefficients:
##                Estimate Std. Error z value Pr(>|z|)  
## (Intercept)     -1.2528     0.8018  -1.562    0.118  
## nhomtuoi150-70   0.6597     0.8466   0.779    0.436  
## nhomtuoi1>70     1.4351     0.8570   1.674    0.094 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 146.99  on 109  degrees of freedom
## AIC: 152.99
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_nhomtuoi1.1), confint(model_nhomtuoi1.1)))
## Waiting for profiling to be done...
##                       OR      2.5 %    97.5 %
## (Intercept)    0.2857143 0.04257115  1.182077
## nhomtuoi150-70 1.9342105 0.42120919 13.778222
## nhomtuoi1>70   4.2000000 0.89583766 30.379652
round(exp(cbind(OR = coef(model_nhomtuoi1.1), confint(model_nhomtuoi1.1))), 2)
## Waiting for profiling to be done...
##                  OR 2.5 % 97.5 %
## (Intercept)    0.29  0.04   1.18
## nhomtuoi150-70 1.93  0.42  13.78
## nhomtuoi1>70   4.20  0.90  30.38
model_gioi.1 <- glm(gpb ~ gioi, data = e1, family = binomial)
summary(model_gioi.1)
## 
## Call:
## glm(formula = gpb ~ gioi, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.5754     0.4167  -1.381    0.167
## gioi          0.3211     0.4694   0.684    0.494
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.88  on 110  degrees of freedom
## AIC: 155.88
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_gioi.1), confint(model_gioi.1)))
## Waiting for profiling to be done...
##                   OR     2.5 %   97.5 %
## (Intercept) 0.562500 0.2380299 1.247758
## gioi        1.378685 0.5580359 3.576569
round(exp(cbind(OR = coef(model_gioi.1), confint(model_gioi.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.56  0.24   1.25
## gioi        1.38  0.56   3.58
e1$hutthuocla2 <- factor(e1$hutthuocla, levels = c(0, 1, 2), labels = c("khong", "co", "Ngung"))
model_htl.2 <- glm(gpb ~ hutthuocla2, data = e1, family = binomial)
summary(model_htl.2)
## 
## Call:
## glm(formula = gpb ~ hutthuocla2, family = binomial, data = e1)
## 
## Coefficients:
##                  Estimate Std. Error z value Pr(>|z|)   
## (Intercept)       -0.7828     0.3018  -2.594  0.00949 **
## hutthuocla2co      0.8661     0.4178   2.073  0.03816 * 
## hutthuocla2Ngung   0.6286     0.6329   0.993  0.32062   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 147.85  on 109  degrees of freedom
## AIC: 153.85
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_htl.2), confint(model_htl.2)))
## Waiting for profiling to be done...
##                         OR     2.5 %    97.5 %
## (Intercept)      0.4571429 0.2464302 0.8116553
## hutthuocla2co    2.3777174 1.0582036 5.4777717
## hutthuocla2Ngung 1.8750000 0.5276059 6.5605628
round(exp(cbind(OR = coef(model_htl.2), confint(model_htl.2))), 2)
## Waiting for profiling to be done...
##                    OR 2.5 % 97.5 %
## (Intercept)      0.46  0.25   0.81
## hutthuocla2co    2.38  1.06   5.48
## hutthuocla2Ngung 1.87  0.53   6.56
model_copd.1 <- glm(gpb ~ copd, data = e1, family = binomial)
summary(model_copd.1)
## 
## Call:
## glm(formula = gpb ~ copd, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.31634    0.20051  -1.578    0.115
## copd        -0.08913    0.67592  -0.132    0.895
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.34  on 110  degrees of freedom
## AIC: 156.34
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_copd.1), confint(model_copd.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7288136 0.4893514 1.076678
## copd        0.9147287 0.2223989 3.398374
round(exp(cbind(OR = coef(model_copd.1), confint(model_copd.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.73  0.49   1.08
## copd        0.91  0.22   3.40
model_lao.1 <- glm(gpb ~ lao, data = e1, family = binomial)
summary(model_lao.1)
## 
## Call:
## glm(formula = gpb ~ lao, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.4595     0.2129  -2.159   0.0309 *
## lao           0.7780     0.5111   1.522   0.1280  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 150.01  on 110  degrees of freedom
## AIC: 154.01
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_lao.1), confint(model_lao.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.6315789 0.4127457 0.9537812
## lao         2.1770833 0.8054137 6.1210460
round(exp(cbind(OR = coef(model_lao.1), confint(model_lao.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.63  0.41   0.95
## lao         2.18  0.81   6.12
model_laomp.1 <- glm(gpb ~ laomp, data = e1, family = binomial)
summary(model_laomp.1)
## 
## Call:
## glm(formula = gpb ~ laomp, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.3302     0.1933  -1.708   0.0876 .
## laomp         0.3302     1.4274   0.231   0.8170  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.31  on 110  degrees of freedom
## AIC: 156.31
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_laomp.1), confint(model_laomp.1)))
## Waiting for profiling to be done...
##                   OR      2.5 %    97.5 %
## (Intercept) 0.718750 0.48959220  1.046897
## laomp       1.391304 0.05408245 35.794499
round(exp(cbind(OR = coef(model_laomp.1), confint(model_laomp.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.72  0.49   1.05
## laomp       1.39  0.05  35.79
model_sot.1 <- glm(gpb ~ sot, data = e1, family = binomial)
summary(model_sot.1)
## 
## Call:
## glm(formula = gpb ~ sot, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept) -0.07796    0.22809  -0.342   0.7325  
## sot         -0.83833    0.43821  -1.913   0.0557 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 148.51  on 110  degrees of freedom
## AIC: 152.51
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_sot.1), confint(model_sot.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.9250000 0.5896666 1.4472458
## sot         0.4324324 0.1769368 0.9987641
round(exp(cbind(OR = coef(model_sot.1), confint(model_sot.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.93  0.59   1.45
## sot         0.43  0.18   1.00
model_met.1 <- glm(gpb ~ met, data = e1, family = binomial)
summary(model_met.1)
## 
## Call:
## glm(formula = gpb ~ met, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2822     0.1953  -1.445    0.148
## met          -1.1041     1.1350  -0.973    0.331
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.23  on 110  degrees of freedom
## AIC: 155.23
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_met.1), confint(model_met.1)))
## Waiting for profiling to be done...
##                    OR      2.5 %   97.5 %
## (Intercept) 0.7540984 0.51191942 1.103230
## met         0.3315217 0.01664777 2.333705
round(exp(cbind(OR = coef(model_met.1), confint(model_met.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.75  0.51   1.10
## met         0.33  0.02   2.33
model_chanan.1 <- glm(gpb ~ chanan, data = e1, family = binomial)
summary(model_chanan.1)
## 
## Call:
## glm(formula = gpb ~ chanan, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.5341     0.2160  -2.473   0.0134 *
## chanan        1.1531     0.5162   2.234   0.0255 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 147.10  on 110  degrees of freedom
## AIC: 151.1
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_chanan.1), confint(model_chanan.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.5862069 0.3802664 0.8896239
## chanan      3.1680672 1.1800137 9.1625380
round(exp(cbind(OR = coef(model_chanan.1), confint(model_chanan.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.59  0.38   0.89
## chanan      3.17  1.18   9.16
model_sutcan.1 <- glm(gpb ~ sutcan, data = e1, family = binomial)
summary(model_sutcan.1)
## 
## Call:
## glm(formula = gpb ~ sutcan, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.5390     0.2746  -1.963   0.0497 *
## sutcan        0.4298     0.3852   1.116   0.2645  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.11  on 110  degrees of freedom
## AIC: 155.11
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_sutcan.1), confint(model_sutcan.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.5833333 0.3351025 0.9897619
## sutcan      1.5369458 0.7244891 3.2948404
round(exp(cbind(OR = coef(model_sutcan.1), confint(model_sutcan.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.58  0.34   0.99
## sutcan      1.54  0.72   3.29
model_hokhan.1 <- glm(gpb ~ hokhan, data = e1, family = binomial)
summary(model_hokhan.1)
## 
## Call:
## glm(formula = gpb ~ hokhan, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.3697     0.2109  -1.753   0.0796 .
## hokhan        0.2644     0.5056   0.523   0.6010  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.09  on 110  degrees of freedom
## AIC: 156.09
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_hokhan.1), confint(model_hokhan.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.6909091 0.4538788 1.040765
## hokhan      1.3026316 0.4754311 3.532063
round(exp(cbind(OR = coef(model_hokhan.1), confint(model_hokhan.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.69  0.45   1.04
## hokhan      1.30  0.48   3.53
model_hodam.1 <- glm(gpb ~ hodam, data = e1, family = binomial)
summary(model_hodam.1)
## 
## Call:
## glm(formula = gpb ~ hodam, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2231     0.3000  -0.744    0.457
## hodam        -0.1699     0.3899  -0.436    0.663
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.17  on 110  degrees of freedom
## AIC: 156.17
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_hodam.1), confint(model_hodam.1)))
## Waiting for profiling to be done...
##                  OR     2.5 %   97.5 %
## (Intercept) 0.80000 0.4396049 1.437147
## hodam       0.84375 0.3921357 1.818037
round(exp(cbind(OR = coef(model_hodam.1), confint(model_hodam.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.80  0.44   1.44
## hodam       0.84  0.39   1.82
model_homau.1 <- glm(gpb ~ homau, data = e1, family = binomial)
summary(model_homau.1)
## 
## Call:
## glm(formula = gpb ~ homau, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.4238     0.2144  -1.977   0.0481 *
## homau         0.5191     0.4867   1.067   0.2861  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.22  on 110  degrees of freedom
## AIC: 155.22
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_homau.1), confint(model_homau.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.6545455 0.4266673 0.9917968
## homau       1.6805556 0.6449437 4.4315823
round(exp(cbind(OR = coef(model_homau.1), confint(model_homau.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.65  0.43   0.99
## homau       1.68  0.64   4.43
model_daunguc.1 <- glm(gpb ~ daunguc, data = e1, family = binomial)
summary(model_daunguc.1)
## 
## Call:
## glm(formula = gpb ~ daunguc, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.7191     0.2798  -2.570   0.0102 *
## daunguc       0.7932     0.3904   2.032   0.0422 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 148.15  on 110  degrees of freedom
## AIC: 152.15
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_daunguc.1), confint(model_daunguc.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.4871795 0.2756922 0.8314211
## daunguc     2.2105263 1.0358224 4.8107166
round(exp(cbind(OR = coef(model_daunguc.1), confint(model_daunguc.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.49  0.28   0.83
## daunguc     2.21  1.04   4.81
model_khotho.1 <- glm(gpb ~ khotho, data = e1, family = binomial)
summary(model_khotho.1)
## 
## Call:
## glm(formula = gpb ~ khotho, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.6931     0.3162  -2.192   0.0284 *
## khotho        0.6035     0.3998   1.510   0.1311  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 150.03  on 110  degrees of freedom
## AIC: 154.03
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_khotho.1), confint(model_khotho.1)))
## Waiting for profiling to be done...
##                   OR     2.5 %   97.5 %
## (Intercept) 0.500000 0.2619361 0.914162
## khotho      1.828571 0.8432246 4.069102
round(exp(cbind(OR = coef(model_khotho.1), confint(model_khotho.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.50  0.26   0.91
## khotho      1.83  0.84   4.07
model_khangiong.1 <- glm(gpb ~ khangiong, data = e1, family = binomial)
summary(model_khangiong.1)
## 
## Call:
## glm(formula = gpb ~ khangiong, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.4212     0.1986  -2.121   0.0339 *
## khangiong     2.0307     1.1133   1.824   0.0682 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 147.75  on 110  degrees of freedom
## AIC: 151.75
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_khangiong.1), confint(model_khangiong.1)))
## Waiting for profiling to be done...
##                   OR    2.5 %      97.5 %
## (Intercept) 0.656250 0.441779   0.9646243
## khangiong   7.619048 1.174862 148.7335249
round(exp(cbind(OR = coef(model_khangiong.1), confint(model_khangiong.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.66  0.44   0.96
## khangiong   7.62  1.17 148.73
model_nuotnghen.1 <- glm(gpb ~ nuotnghen, data = e1, family = binomial)
summary(model_nuotnghen.1)
## 
## Call:
## glm(formula = gpb ~ nuotnghen, family = binomial, data = e1)
## 
## Coefficients:
##              Estimate Std. Error z value Pr(>|z|)  
## (Intercept)   -0.4367     0.1980  -2.206   0.0274 *
## nuotnghen     17.0028  1073.1090   0.016   0.9874  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 143.35  on 110  degrees of freedom
## AIC: 147.35
## 
## Number of Fisher Scoring iterations: 15
model_sohach.1 <- glm(gpb ~ sohach, data = e1, family = binomial)
summary(model_sohach.1)
## 
## Call:
## glm(formula = gpb ~ sohach, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.4383     0.2028  -2.161   0.0307 *
## sohach        1.2856     0.7192   1.787   0.0739 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 148.84  on 110  degrees of freedom
## AIC: 152.84
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_sohach.1), confint(model_sohach.1)))
## Waiting for profiling to be done...
##                    OR    2.5 %     97.5 %
## (Intercept) 0.6451613 0.430509  0.9558605
## sohach      3.6166667 0.945904 17.5416455
round(exp(cbind(OR = coef(model_sohach.1), confint(model_sohach.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.65  0.43   0.96
## sohach      3.62  0.95  17.54
model_hc3giam.1<- glm(gpb ~ hc3giam, data = e1, family = binomial)
summary(model_hc3giam.1)
## 
## Call:
## glm(formula = gpb ~ hc3giam, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.34033    0.21508  -1.582    0.114
## hc3giam      0.07796    0.47242   0.165    0.869
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.33  on 110  degrees of freedom
## AIC: 156.33
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_hc3giam.1), confint(model_hc3giam.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7115385 0.4637209 1.080884
## hc3giam     1.0810811 0.4199047 2.722497
round(exp(cbind(OR = coef(model_hc3giam.1), confint(model_hc3giam.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.71  0.46   1.08
## hc3giam     1.08  0.42   2.72
model_tn.1<- glm(gpb ~ tn, data = e1, family = binomial)
summary(model_tn.1)
## 
## Call:
## glm(formula = gpb ~ tn, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)   
## (Intercept)  -0.6745     0.2364  -2.853  0.00434 **
## tn            1.1853     0.4350   2.725  0.00644 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 144.64  on 110  degrees of freedom
## AIC: 148.64
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_tn.1), confint(model_tn.1)))
## Waiting for profiling to be done...
##                   OR     2.5 %    97.5 %
## (Intercept) 0.509434 0.3160877 0.8020456
## tn          3.271605 1.4138525 7.8590962
round(exp(cbind(OR = coef(model_tn.1), confint(model_tn.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.51  0.32   0.80
## tn          3.27  1.41   7.86
model_nt.1<- glm(gpb ~ nt, data = e1, family = binomial)
summary(model_nt.1)
## 
## Call:
## glm(formula = gpb ~ nt, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept) -0.05264    0.22950  -0.229   0.8186  
## nt          -0.90287    0.43718  -2.065   0.0389 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 147.85  on 110  degrees of freedom
## AIC: 151.85
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_nt.1), confint(model_nt.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.9487179 0.6033565 1.4891018
## nt          0.4054054 0.1661177 0.9337894
round(exp(cbind(OR = coef(model_nt.1), confint(model_nt.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.95  0.60   1.49
## nt          0.41  0.17   0.93
model_tt.1<- glm(gpb ~ tt, data = e1, family = binomial)
summary(model_tt.1)
## 
## Call:
## glm(formula = gpb ~ tt, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.3589     0.1965  -1.827   0.0677 .
## tt            0.7644     0.9338   0.819   0.4130  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.67  on 110  degrees of freedom
## AIC: 155.67
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_tt.1), confint(model_tt.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.6984127 0.4725376  1.02319
## tt          2.1477273 0.3425033 16.81516
round(exp(cbind(OR = coef(model_tt.1), confint(model_tt.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.70  0.47   1.02
## tt          2.15  0.34  16.82
model_khonggn.1<- glm(gpb ~ khonggn, data = e1, family = binomial)
summary(model_khonggn.1)
## 
## Call:
## glm(formula = gpb ~ khonggn, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.4372     0.2304  -1.898   0.0578 .
## khonggn       0.3766     0.4176   0.902   0.3672  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.55  on 110  degrees of freedom
## AIC: 155.55
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_khonggn.1), confint(model_khonggn.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.6458333 0.4073263 1.009030
## khonggn     1.4573055 0.6403680 3.319837
round(exp(cbind(OR = coef(model_khonggn.1), confint(model_khonggn.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.65  0.41   1.01
## khonggn     1.46  0.64   3.32
model_kinhmo.1<- glm(gpb ~ kinhmo, data = e1, family = binomial)
summary(model_kinhmo.1)
## 
## Call:
## glm(formula = gpb ~ kinhmo, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.4055     0.2152  -1.884   0.0595 .
## kinhmo        0.4055     0.4776   0.849   0.3959  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.64  on 110  degrees of freedom
## AIC: 155.64
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_kinhmo.1), confint(model_kinhmo.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.6666667 0.4339887 1.011951
## kinhmo      1.5000000 0.5836481 3.863890
round(exp(cbind(OR = coef(model_kinhmo.1), confint(model_kinhmo.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.67  0.43   1.01
## kinhmo      1.50  0.58   3.86
model_hang.1<- glm(gpb ~ hang, data = e1, family = binomial)
summary(model_hang.1)
## 
## Call:
## glm(formula = gpb ~ hang, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.1787     0.1998  -0.894   0.3711  
## hang         -2.1239     1.0675  -1.990   0.0466 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 145.91  on 110  degrees of freedom
## AIC: 149.91
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_hang.1), confint(model_hang.1)))
## Waiting for profiling to be done...
##                    OR       2.5 %    97.5 %
## (Intercept) 0.8363636 0.563353289 1.2360059
## hang        0.1195652 0.006395857 0.6589616
round(exp(cbind(OR = coef(model_hang.1), confint(model_hang.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.84  0.56   1.24
## hang        0.12  0.01   0.66
model_not.1<- glm(gpb ~ notmo, data = e1, family = binomial)
summary(model_not.1)
## 
## Call:
## glm(formula = gpb ~ notmo, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept) -0.07411    0.22237  -0.333   0.7389  
## notmo       -0.98194    0.46683  -2.103   0.0354 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 147.58  on 110  degrees of freedom
## AIC: 151.58
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_not.1), confint(model_not.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.9285714 0.5987744 1.4366219
## notmo       0.3745819 0.1426495 0.9059756
round(exp(cbind(OR = coef(model_not.1), confint(model_not.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.93  0.60   1.44
## notmo       0.37  0.14   0.91
model_dam.1<- glm(gpb ~ dam, data = e1, family = binomial)
summary(model_dam.1)
## 
## Call:
## glm(formula = gpb ~ dam, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)   -17.57    1097.25  -0.016    0.987
## dam            17.46    1097.25   0.016    0.987
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 136.99  on 110  degrees of freedom
## AIC: 140.99
## 
## Number of Fisher Scoring iterations: 16
exp(cbind(OR = coef(model_dam.1), confint(model_dam.1)))
## Waiting for profiling to be done...
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
##                       OR        2.5 %       97.5 %
## (Intercept) 2.350463e-08           NA 2.149555e+25
## dam         3.845396e+07 5.337581e-26           NA
round(exp(cbind(OR = coef(model_dam.1), confint(model_dam.1))), 2)
## Waiting for profiling to be done...
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
##                   OR 2.5 %       97.5 %
## (Intercept)        0    NA 2.149555e+25
## dam         38453965     0           NA
model_dongdac.1<- glm(gpb ~ dongdac, data = e1, family = binomial)
summary(model_dongdac.1)
## 
## Call:
## glm(formula = gpb ~ dongdac, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2624     0.2428  -1.080    0.280
## dongdac      -0.1625     0.3953  -0.411    0.681
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.19  on 110  degrees of freedom
## AIC: 156.19
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_dongdac.1), confint(model_dongdac.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7692308 0.4744653 1.234812
## dongdac     0.8500000 0.3879816 1.838557
round(exp(cbind(OR = coef(model_dongdac.1), confint(model_dongdac.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.77  0.47   1.23
## dongdac     0.85  0.39   1.84
model_xep.1<- glm(gpb ~ xep, data = e1, family = binomial)
summary(model_xep.1)
## 
## Call:
## glm(formula = gpb ~ xep, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.3697     0.2109  -1.753   0.0796 .
## xep           0.2644     0.5056   0.523   0.6010  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.09  on 110  degrees of freedom
## AIC: 156.09
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_xep.1), confint(model_xep.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.6909091 0.4538788 1.040765
## xep         1.3026316 0.4754311 3.532063
round(exp(cbind(OR = coef(model_xep.1), confint(model_xep.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.69  0.45   1.04
## xep         1.30  0.48   3.53
model_tdmp.1<- glm(gpb ~ tdmp, data = e1, family = binomial)
summary(model_tdmp.1)
## 
## Call:
## glm(formula = gpb ~ tdmp, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.28768    0.25459  -1.130    0.258
## tdmp        -0.08388    0.38639  -0.217    0.828
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.31  on 110  degrees of freedom
## AIC: 156.31
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_tdmp.1), confint(model_tdmp.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7500000 0.4514642 1.231149
## tdmp        0.9195402 0.4287466 1.959928
round(exp(cbind(OR = coef(model_tdmp.1), confint(model_tdmp.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.75  0.45   1.23
## tdmp        0.92  0.43   1.96
model_trenp.1<- glm(gpb ~ trenp, data = e1, family = binomial)
summary(model_trenp.1)
## 
## Call:
## glm(formula = gpb ~ trenp, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)   
## (Intercept)   0.1643     0.2569   0.639  0.52252   
## trenp        -1.1362     0.4056  -2.802  0.00509 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 144.10  on 110  degrees of freedom
## AIC: 148.1
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_trenp.1), confint(model_trenp.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 1.1785714 0.7128543 1.9622026
## trenp       0.3210483 0.1418272 0.7001621
round(exp(cbind(OR = coef(model_trenp.1), confint(model_trenp.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 1.18  0.71   1.96
## trenp       0.32  0.14   0.70
model_giuap.1<- glm(gpb ~ giuap, data = e1, family = binomial)
summary(model_giuap.1)
## 
## Call:
## glm(formula = gpb ~ giuap, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2877     0.2205  -1.305    0.192
## giuap        -0.1476     0.4454  -0.331    0.740
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.25  on 110  degrees of freedom
## AIC: 156.25
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_giuap.1), confint(model_giuap.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7500000 0.4838458 1.152206
## giuap       0.8627451 0.3527618 2.048359
round(exp(cbind(OR = coef(model_giuap.1), confint(model_giuap.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.75  0.48   1.15
## giuap       0.86  0.35   2.05
model_duoip.1<- glm(gpb ~ duoip, data = e1, family = binomial)
summary(model_duoip.1)
## 
## Call:
## glm(formula = gpb ~ duoip, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.3295     0.2048  -1.609    0.108
## duoip         0.0418     0.5776   0.072    0.942
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.35  on 110  degrees of freedom
## AIC: 156.35
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_duoip.1), confint(model_duoip.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7192982 0.4787336 1.071262
## duoip       1.0426829 0.3214429 3.225456
round(exp(cbind(OR = coef(model_duoip.1), confint(model_duoip.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.72  0.48   1.07
## duoip       1.04  0.32   3.23
model_trent.1<- glm(gpb ~ trent, data = e1, family = binomial)
summary(model_trent.1)
## 
## Call:
## glm(formula = gpb ~ trent, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.6131     0.2435  -2.518   0.0118 *
## trent         0.8244     0.4071   2.025   0.0429 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 148.20  on 110  degrees of freedom
## AIC: 152.2
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_trent.1), confint(model_trent.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.5416667 0.3315188 0.8651394
## trent       2.2805430 1.0323278 5.1261785
round(exp(cbind(OR = coef(model_trent.1), confint(model_trent.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.54  0.33   0.87
## trent       2.28  1.03   5.13
model_duoit.1<- glm(gpb ~ duoit, data = e1, family = binomial)
summary(model_duoit.1)
## 
## Call:
## glm(formula = gpb ~ duoit, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2429     0.2112  -1.150    0.250
## duoit        -0.4502     0.5088  -0.885    0.376
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.55  on 110  degrees of freedom
## AIC: 155.55
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_duoit.1), confint(model_duoit.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7843137 0.5158871 1.184156
## duoit       0.6375000 0.2233456 1.684950
round(exp(cbind(OR = coef(model_duoit.1), confint(model_duoit.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.78  0.52   1.18
## duoit       0.64  0.22   1.68
model_nho3.1<- glm(gpb ~ nho3, data = e1, family = binomial)
summary(model_nho3.1)
## 
## Call:
## glm(formula = gpb ~ nho3, family = binomial, data = e1)
## 
## Coefficients:
##              Estimate Std. Error z value Pr(>|z|)
## (Intercept)   -0.1011     0.2013  -0.502    0.615
## nho3         -17.4650  1097.2470  -0.016    0.987
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 136.99  on 110  degrees of freedom
## AIC: 140.99
## 
## Number of Fisher Scoring iterations: 16
exp(cbind(OR = coef(model_nho3.1), confint(model_nho3.1)))
## Waiting for profiling to be done...
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
##                       OR     2.5 %       97.5 %
## (Intercept) 9.038462e-01 0.6076408 1.340905e+00
## nho3        2.600512e-08        NA 1.873509e+25
round(exp(cbind(OR = coef(model_nho3.1), confint(model_nho3.1))), 2)
## Waiting for profiling to be done...
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
##              OR 2.5 %       97.5 %
## (Intercept) 0.9  0.61 1.340000e+00
## nho3        0.0    NA 1.873509e+25
model_tu35.1<- glm(gpb ~ tu35, data = e1, family = binomial)
summary(model_tu35.1)
## 
## Call:
## glm(formula = gpb ~ tu35, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.3830     0.2368  -1.618    0.106
## tu35          0.1717     0.4031   0.426    0.670
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.18  on 110  degrees of freedom
## AIC: 156.18
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_tu35.1), confint(model_tu35.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.6818182 0.4248539 1.079343
## tu35        1.1873016 0.5356103 2.618985
round(exp(cbind(OR = coef(model_tu35.1), confint(model_tu35.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.68  0.42   1.08
## tu35        1.19  0.54   2.62
model_tu57.1<- glm(gpb ~ tu57, data = e1, family = binomial)
summary(model_tu57.1)
## 
## Call:
## glm(formula = gpb ~ tu57, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.4738     0.2217  -2.137   0.0326 *
## tu57          0.6279     0.4516   1.391   0.1644  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 150.42  on 110  degrees of freedom
## AIC: 154.42
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_tu57.1), confint(model_tu57.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.6226415 0.3995115 0.9562005
## tu57        1.8737374 0.7740147 4.6057053
round(exp(cbind(OR = coef(model_tu57.1), confint(model_tu57.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.62  0.40   0.96
## tu57        1.87  0.77   4.61
model_lon7.1<- glm(gpb ~ lon7, data = e1, family = binomial)
summary(model_lon7.1)
## 
## Call:
## glm(formula = gpb ~ lon7, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.3947     0.2324  -1.698   0.0894 .
## lon7          0.2228     0.4113   0.542   0.5880  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.07  on 110  degrees of freedom
## AIC: 156.07
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_lon7.1), confint(model_lon7.1)))
## Waiting for profiling to be done...
##                   OR     2.5 %   97.5 %
## (Intercept) 0.673913 0.4236365 1.057612
## lon7        1.249576 0.5546321 2.802534
round(exp(cbind(OR = coef(model_lon7.1), confint(model_lon7.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.67  0.42   1.06
## lon7        1.25  0.55   2.80
e1$bo3 <- factor(e1$bo, levels = c(1, 2, 3), labels = c("tron_deu", "da_cung", "tua_gai"))
# Tạo biến mới bo4 từ biến bo
e1$bo4 <- factor(e1$bo, levels = c(1, 2, 3), labels = c(0, 1, 2))

model_bo4.1 <- glm(gpb ~ bo4, data = e1, family = binomial)
summary(model_bo4.1)
## 
## Call:
## glm(formula = gpb ~ bo4, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -1.5041     0.7817  -1.924   0.0544 .
## bo41          0.8109     0.8333   0.973   0.3305  
## bo42          1.8042     0.8356   2.159   0.0308 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 143.28  on 109  degrees of freedom
## AIC: 149.28
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_bo4.1), confint(model_bo4.1)))
## Waiting for profiling to be done...
##                    OR      2.5 %     97.5 %
## (Intercept) 0.2222222 0.03387787  0.8622328
## bo41        2.2500000 0.51142764 15.7837538
## bo42        6.0750000 1.37999207 42.8078397
round(exp(cbind(OR = coef(model_bo4.1), confint(model_bo4.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.22  0.03   0.86
## bo41        2.25  0.51  15.78
## bo42        6.07  1.38  42.81
model_hach.1 <- glm(gpb ~ hach, data = e1, family = binomial)
summary(model_hach.1)
## 
## Call:
## glm(formula = gpb ~ hach, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.5021     0.2642  -1.901   0.0574 .
## hach          0.3843     0.3854   0.997   0.3186  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.36  on 110  degrees of freedom
## AIC: 155.36
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_hach.1), confint(model_hach.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.6052632 0.3556082 1.007622
## hach        1.4685990 0.6908658 3.144383
round(exp(cbind(OR = coef(model_hach.1), confint(model_hach.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.61  0.36   1.01
## hach        1.47  0.69   3.14
model_khongxl.1 <- glm(gpb ~ khongxl, data = e1, family = binomial)
summary(model_khongxl.1)
## 
## Call:
## glm(formula = gpb ~ khongxl, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2048     0.2872  -0.713    0.476
## khongxl      -0.2139     0.3858  -0.555    0.579
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.05  on 110  degrees of freedom
## AIC: 156.05
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_khongxl.1), confint(model_khongxl.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.8148148 0.4598224 1.428246
## khongxl     0.8074163 0.3778081 1.722468
round(exp(cbind(OR = coef(model_khongxl.1), confint(model_khongxl.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.81  0.46   1.43
## khongxl     0.81  0.38   1.72
model_xlmp.1 <- glm(gpb ~ xlmp, data = e1, family = binomial)
summary(model_xlmp.1)
## 
## Call:
## glm(formula = gpb ~ xlmp, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.4520     0.2162  -2.090   0.0366 *
## xlmp          0.6343     0.4797   1.322   0.1860  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 150.60  on 110  degrees of freedom
## AIC: 154.6
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_xlmp.1), confint(model_xlmp.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.6363636 0.4130902 0.9672841
## xlmp        1.8857143 0.7372032 4.9212646
round(exp(cbind(OR = coef(model_xlmp.1), confint(model_xlmp.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.64  0.41   0.97
## xlmp        1.89  0.74   4.92
model_xltn.1 <- glm(gpb ~ xltn, data = e1, family = binomial)
summary(model_xltn.1)
## 
## Call:
## glm(formula = gpb ~ xltn, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.3640     0.2033  -1.790   0.0734 .
## xltn          0.3640     0.6121   0.595   0.5521  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.01  on 110  degrees of freedom
## AIC: 156.01
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_xltn.1), confint(model_xltn.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.6949153 0.4636632 1.031574
## xltn        1.4390244 0.4226504 4.904081
round(exp(cbind(OR = coef(model_xltn.1), confint(model_xltn.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.69  0.46   1.03
## xltn        1.44  0.42   4.90
model_xltt.1 <- glm(gpb ~ xltt, data = e1, family = binomial)
summary(model_xltt.1)
## 
## Call:
## glm(formula = gpb ~ xltt, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.5500     0.2293  -2.399   0.0164 *
## xltt          0.8183     0.4339   1.886   0.0593 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 148.75  on 110  degrees of freedom
## AIC: 152.75
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_xltt.1), confint(model_xltt.1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.5769231 0.3640605 0.8977806
## xltt        2.2666667 0.9739394 5.3925165
round(exp(cbind(OR = coef(model_xltt.1), confint(model_xltt.1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.58  0.36   0.90
## xltt        2.27  0.97   5.39
model_ca1 <- glm(gpb ~ carina, data = e1, family = binomial)
summary(model_ca1)
## 
## Call:
## glm(formula = gpb ~ carina, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.3522     0.1945  -1.810   0.0702 .
## carina        1.0454     1.2401   0.843   0.3992  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.60  on 110  degrees of freedom
## AIC: 155.6
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_ca1), confint(model_ca1)))
## Waiting for profiling to be done...
##                   OR     2.5 %    97.5 %
## (Intercept) 0.703125 0.4776211  1.026346
## carina      2.844444 0.2648030 62.299279
round(exp(cbind(OR = coef(model_ca1), confint(model_ca1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.70  0.48   1.03
## carina      2.84  0.26  62.30
model_gocp1 <- glm(gpb ~ gocp, data = e1, family = binomial)
summary(model_gocp1)
## 
## Call:
## glm(formula = gpb ~ gocp, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2992     0.2012  -1.487    0.137
## gocp         -0.2604     0.6583  -0.396    0.692
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.20  on 110  degrees of freedom
## AIC: 156.2
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_gocp1), confint(model_gocp1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7413793 0.4971663 1.097025
## gocp        0.7707641 0.1918181 2.719771
round(exp(cbind(OR = coef(model_gocp1), confint(model_gocp1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.74  0.50   1.10
## gocp        0.77  0.19   2.72
model_goct1 <- glm(gpb ~ goct, data = e1, family = binomial)
summary(model_goct1)
## 
## Call:
## glm(formula = gpb ~ goct, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.4473     0.2050  -2.182   0.0291 *
## goct          1.1405     0.6458   1.766   0.0774 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 149.03  on 110  degrees of freedom
## AIC: 153.03
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_goct1), confint(model_goct1)))
## Waiting for profiling to be done...
##                    OR     2.5 %     97.5 %
## (Intercept) 0.6393443 0.4246526  0.9511848
## goct        3.1282051 0.9206413 12.3729094
round(exp(cbind(OR = coef(model_goct1), confint(model_goct1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.64  0.42   0.95
## goct        3.13  0.92  12.37
model_pqtp1<- glm(gpb ~ pqtp, data = e1, family = binomial)
summary(model_pqtp1)
## 
## Call:
## glm(formula = gpb ~ pqtp, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.1591     0.2528  -0.629    0.529
## pqtp         -0.3846     0.3895  -0.987    0.323
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.38  on 110  degrees of freedom
## AIC: 155.38
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_pqtp1), confint(model_pqtp1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.8529412 0.5166872 1.398843
## pqtp        0.6807564 0.3140861 1.453899
round(exp(cbind(OR = coef(model_pqtp1), confint(model_pqtp1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.85  0.52   1.40
## pqtp        0.68  0.31   1.45
model_pqgp1 <- glm(gpb ~ pqgp, data = e1, family = binomial)
summary(model_pqgp1)
## 
## Call:
## glm(formula = gpb ~ pqgp, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2822     0.1953  -1.445    0.148
## pqgp         -1.1041     1.1350  -0.973    0.331
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.23  on 110  degrees of freedom
## AIC: 155.23
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_pqgp1), confint(model_pqgp1)))
## Waiting for profiling to be done...
##                    OR      2.5 %   97.5 %
## (Intercept) 0.7540984 0.51191942 1.103230
## pqgp        0.3315217 0.01664777 2.333705
round(exp(cbind(OR = coef(model_pqgp1), confint(model_pqgp1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.75  0.51   1.10
## pqgp        0.33  0.02   2.33
model_pqtg1 <- glm(gpb ~ pqtg, data = e1, family = binomial)
summary(model_pqtg1)
## 
## Call:
## glm(formula = gpb ~ pqtg, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2683     0.2127  -1.261    0.207
## pqtg         -0.2914     0.4916  -0.593    0.553
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.00  on 110  degrees of freedom
## AIC: 156
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_pqtg1), confint(model_pqtg1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7647059 0.5012773 1.157540
## pqtg        0.7472527 0.2742059 1.925474
round(exp(cbind(OR = coef(model_pqtg1), confint(model_pqtg1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.76  0.50   1.16
## pqtg        0.75  0.27   1.93
model_pqdp1 <- glm(gpb ~ pqdp, data = e1, family = binomial)
summary(model_pqdp1)
## 
## Call:
## glm(formula = gpb ~ pqdp, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2938     0.2063  -1.424    0.155
## pqdp         -0.2171     0.5561  -0.390    0.696
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 152.21  on 110  degrees of freedom
## AIC: 156.21
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_pqdp1), confint(model_pqdp1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7454545 0.4948242 1.114088
## pqdp        0.8048780 0.2558628 2.348363
round(exp(cbind(OR = coef(model_pqdp1), confint(model_pqdp1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.75  0.49   1.11
## pqdp        0.80  0.26   2.35
model_pqtt1 <- glm(gpb ~ pqtt, data = e1, family = binomial)
summary(model_pqtt1)
## 
## Call:
## glm(formula = gpb ~ pqtt, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)   
## (Intercept)  -0.6242     0.2263  -2.759  0.00580 **
## pqtt          1.2601     0.4702   2.680  0.00737 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 144.78  on 110  degrees of freedom
## AIC: 148.78
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_pqtt1), confint(model_pqtt1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 0.5357143 0.3397819 0.8279557
## pqtt        3.5259259 1.4317904 9.1942009
round(exp(cbind(OR = coef(model_pqtt1), confint(model_pqtt1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.54  0.34   0.83
## pqtt        3.53  1.43   9.19
model_pqdt1 <- glm(gpb ~ pqdt, data = e1, family = binomial)
summary(model_pqdt1)
## 
## Call:
## glm(formula = gpb ~ pqdt, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2231     0.2121  -1.052    0.293
## pqdt         -0.5390     0.5045  -1.068    0.285
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.17  on 110  degrees of freedom
## AIC: 155.17
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_pqdt1), confint(model_pqdt1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.8000000 0.5253731 1.210362
## pqdt        0.5833333 0.2056770 1.524694
round(exp(cbind(OR = coef(model_pqdt1), confint(model_pqdt1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.80  0.53   1.21
## pqdt        0.58  0.21   1.52
model_xhuyet1 <- glm(gpb ~ xhuyet, data = e1, family = binomial)
summary(model_xhuyet1)
## 
## Call:
## glm(formula = gpb ~ xhuyet, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.4055     0.2214  -1.831   0.0671 .
## xhuyet        0.3314     0.4443   0.746   0.4558  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.80  on 110  degrees of freedom
## AIC: 155.8
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_xhuyet1), confint(model_xhuyet1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.6666667 0.4285028 1.024185
## xhuyet      1.3928571 0.5788131 3.343733
round(exp(cbind(OR = coef(model_xhuyet1), confint(model_xhuyet1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.67  0.43   1.02
## xhuyet      1.39  0.58   3.34
model_thamnhiem1 <- glm(gpb ~ thamnhiem, data = e1, family = binomial)
summary(model_thamnhiem1)
## 
## Call:
## glm(formula = gpb ~ thamnhiem, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept)   0.4700     0.4031   1.166   0.2436  
## thamnhiem    -1.0433     0.4615  -2.261   0.0238 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 147.08  on 110  degrees of freedom
## AIC: 151.08
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_thamnhiem1), confint(model_thamnhiem1)))
## Waiting for profiling to be done...
##                    OR     2.5 %    97.5 %
## (Intercept) 1.6000000 0.7362311 3.6509338
## thamnhiem   0.3522727 0.1386214 0.8586579
round(exp(cbind(OR = coef(model_thamnhiem1), confint(model_thamnhiem1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 1.60  0.74   3.65
## thamnhiem   0.35  0.14   0.86
model_noilong <- glm (gpb ~ noilong, data=e1, family = binomial)
summary(model_noilong)
## 
## Call:
## glm(formula = gpb ~ noilong, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)    
## (Intercept)  -1.1787     0.3301  -3.570 0.000356 ***
## noilong       1.4759     0.4195   3.518 0.000435 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 138.88  on 110  degrees of freedom
## AIC: 142.88
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_noilong), confint(model_noilong)))
## Waiting for profiling to be done...
##                    OR     2.5 %     97.5 %
## (Intercept) 0.3076923 0.1542453  0.5696271
## noilong     4.3750000 1.9640835 10.2575700
round(exp(cbind(OR = coef(model_noilong), confint(model_noilong))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.31  0.15   0.57
## noilong     4.37  1.96  10.26
model_usui1 <- glm(gpb ~ usui, data = e1, family = binomial)
summary(model_usui1)
## 
## Call:
## glm(formula = gpb ~ usui, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)   
## (Intercept)  -0.7621     0.2643  -2.884  0.00393 **
## usui          1.0245     0.3979   2.575  0.01003 * 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 145.55  on 110  degrees of freedom
## AIC: 149.55
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_usui1), confint(model_usui1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.4666667 0.2726621 0.772966
## usui        2.7857143 1.2883247 6.161661
round(exp(cbind(OR = coef(model_usui1), confint(model_usui1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.47  0.27   0.77
## usui        2.79  1.29   6.16
model_chenep1 <- glm(gpb ~ chenep, data = e1, family = binomial)
summary(model_chenep1)
## 
## Call:
## glm(formula = gpb ~ chenep, family = binomial, data = e1)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)
## (Intercept)  -0.2461     0.2036  -1.209    0.227
## chenep       -0.6702     0.6256  -1.071    0.284
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 151.14  on 110  degrees of freedom
## AIC: 155.14
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_chenep1), confint(model_chenep1)))
## Waiting for profiling to be done...
##                    OR     2.5 %   97.5 %
## (Intercept) 0.7818182 0.5221964 1.162895
## chenep      0.5116279 0.1330431 1.645438
round(exp(cbind(OR = coef(model_chenep1), confint(model_chenep1))), 2)
## Waiting for profiling to be done...
##               OR 2.5 % 97.5 %
## (Intercept) 0.78  0.52   1.16
## chenep      0.51  0.13   1.65
# Tạo biến mới vitri3
e1$vitri3 <- ifelse(e1$vtri == 1, 1, 0)

# Chuyển thành factor với nhãn rõ ràng
e1$vitri3 <- factor(e1$vitri3, levels = c(0, 1),
                      labels = c("Niêm mạc/dưới niêm", "Khối u"))
model_vtri1 <- glm(gpb ~ vitri3, data = e1, family = binomial)
summary(model_vtri1)
## 
## Call:
## glm(formula = gpb ~ vitri3, family = binomial, data = e1)
## 
## Coefficients:
##              Estimate Std. Error z value Pr(>|z|)    
## (Intercept)   -0.8535     0.2490  -3.428 0.000609 ***
## vitri3Khối u   1.6336     0.4411   3.704 0.000213 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 137.48  on 110  degrees of freedom
## AIC: 141.48
## 
## Number of Fisher Scoring iterations: 4
exp(cbind(OR = coef(model_vtri1), confint(model_vtri1)))
## Waiting for profiling to be done...
##                     OR     2.5 %    97.5 %
## (Intercept)  0.4259259 0.2566226  0.684543
## vitri3Khối u 5.1225296 2.2054349 12.555950
round(exp(cbind(OR = coef(model_vtri1), confint(model_vtri1))), 2)
## Waiting for profiling to be done...
##                OR 2.5 % 97.5 %
## (Intercept)  0.43  0.26   0.68
## vitri3Khối u 5.12  2.21  12.56
e1$sluong1 <- factor(e1$soluong, levels = c(1, 2, 3, 4, 5))
model_sluong1 <- glm(gpb ~ sluong1, data = e1, family = binomial)
summary(model_sluong1)
## 
## Call:
## glm(formula = gpb ~ sluong1, family = binomial, data = e1)
## 
## Coefficients:
##               Estimate Std. Error z value Pr(>|z|)
## (Intercept)   -0.69315    1.22474  -0.566    0.571
## sluong12       0.69315    1.25974   0.550    0.582
## sluong13      -0.02899    1.25992  -0.023    0.982
## sluong14       1.09861    1.38444   0.794    0.427
## sluong15     -14.87292 1455.39805  -0.010    0.992
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance: 146.77  on 107  degrees of freedom
## AIC: 156.77
## 
## Number of Fisher Scoring iterations: 14
exp(cbind(OR = coef(model_sluong1), confint(model_sluong1)))
## Waiting for profiling to be done...
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: algorithm did not converge
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
##                       OR      2.5 %        97.5 %
## (Intercept) 5.000000e-01 0.02323545  5.220163e+00
## sluong12    2.000000e+00 0.17944219  4.481807e+01
## sluong13    9.714286e-01 0.08704157  2.176502e+01
## sluong14    3.000000e+00 0.21372310  7.876075e+01
## sluong15    3.473541e-07         NA 9.736723e+101
round(exp(cbind(OR = coef(model_sluong1), confint(model_sluong1))), 2)
## Waiting for profiling to be done...
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: algorithm did not converge
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
##               OR 2.5 %        97.5 %
## (Intercept) 0.50  0.02  5.220000e+00
## sluong12    2.00  0.18  4.482000e+01
## sluong13    0.97  0.09  2.177000e+01
## sluong14    3.00  0.21  7.876000e+01
## sluong15    0.00    NA 9.736723e+101
model_full1 <- glm(
  gpb ~ hutthuocla2 + chanan   + daunguc +
    tn + nt + hang + notmo + trenp + trent  + bo4 
   + xltt + pqtt + usui + noilong + vitri3,
  data = e1,
  family = binomial
)
summary(model_full1)
## 
## Call:
## glm(formula = gpb ~ hutthuocla2 + chanan + daunguc + tn + nt + 
##     hang + notmo + trenp + trent + bo4 + xltt + pqtt + usui + 
##     noilong + vitri3, family = binomial, data = e1)
## 
## Coefficients:
##                  Estimate Std. Error z value Pr(>|z|)   
## (Intercept)       -4.4285     1.4969  -2.958  0.00309 **
## hutthuocla2co      1.0586     0.5970   1.773  0.07620 . 
## hutthuocla2Ngung   1.3235     0.9646   1.372  0.17002   
## chanan             0.8013     0.6219   1.288  0.19763   
## daunguc            0.8056     0.5437   1.482  0.13841   
## tn                 0.6541     0.6452   1.014  0.31064   
## nt                -0.2837     0.6299  -0.450  0.65246   
## hang              -2.4922     1.2392  -2.011  0.04431 * 
## notmo             -0.9076     0.6764  -1.342  0.17968   
## trenp             -0.5973     0.6721  -0.889  0.37416   
## trent              0.7382     0.7660   0.964  0.33517   
## bo41               1.9450     1.3110   1.484  0.13792   
## bo42               2.2943     1.2923   1.775  0.07583 . 
## xltt               0.3118     0.6267   0.498  0.61881   
## pqtt               0.9327     0.7164   1.302  0.19291   
## usui              -0.7136     0.9875  -0.723  0.46992   
## noilong            1.2173     0.8212   1.482  0.13826   
## vitri3Khối u       1.3872     0.8761   1.583  0.11332   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.360  on 111  degrees of freedom
## Residual deviance:  96.431  on  94  degrees of freedom
## AIC: 132.43
## 
## Number of Fisher Scoring iterations: 5
model_step <- step(glm(gpb ~ hutthuocla2 + chanan   + daunguc +
    tn + nt + hang + notmo + trenp + trent  + bo4 
   + xltt + pqtt + usui + noilong + vitri3, family = binomial, 
    data = e1),  scope = formula(model_full1), direction = "forward")
## Start:  AIC=132.43
## gpb ~ hutthuocla2 + chanan + daunguc + tn + nt + hang + notmo + 
##     trenp + trent + bo4 + xltt + pqtt + usui + noilong + vitri3
summary(model_step)
## 
## Call:
## glm(formula = gpb ~ hutthuocla2 + chanan + daunguc + tn + nt + 
##     hang + notmo + trenp + trent + bo4 + xltt + pqtt + usui + 
##     noilong + vitri3, family = binomial, data = e1)
## 
## Coefficients:
##                  Estimate Std. Error z value Pr(>|z|)   
## (Intercept)       -4.4285     1.4969  -2.958  0.00309 **
## hutthuocla2co      1.0586     0.5970   1.773  0.07620 . 
## hutthuocla2Ngung   1.3235     0.9646   1.372  0.17002   
## chanan             0.8013     0.6219   1.288  0.19763   
## daunguc            0.8056     0.5437   1.482  0.13841   
## tn                 0.6541     0.6452   1.014  0.31064   
## nt                -0.2837     0.6299  -0.450  0.65246   
## hang              -2.4922     1.2392  -2.011  0.04431 * 
## notmo             -0.9076     0.6764  -1.342  0.17968   
## trenp             -0.5973     0.6721  -0.889  0.37416   
## trent              0.7382     0.7660   0.964  0.33517   
## bo41               1.9450     1.3110   1.484  0.13792   
## bo42               2.2943     1.2923   1.775  0.07583 . 
## xltt               0.3118     0.6267   0.498  0.61881   
## pqtt               0.9327     0.7164   1.302  0.19291   
## usui              -0.7136     0.9875  -0.723  0.46992   
## noilong            1.2173     0.8212   1.482  0.13826   
## vitri3Khối u       1.3872     0.8761   1.583  0.11332   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.360  on 111  degrees of freedom
## Residual deviance:  96.431  on  94  degrees of freedom
## AIC: 132.43
## 
## Number of Fisher Scoring iterations: 5
#Chạy bằng pp Stepwise
# Model khởi đầu: chỉ có intercept
model_start <- glm(gpb ~ 1, family = binomial, data = e1)

# Model đầy đủ: chứa tất cả biến độc lập
model_full <- glm(gpb ~ hutthuocla2 + chanan   + daunguc +
    tn + nt + hang + notmo + trenp + trent  + bo4 
   + xltt + pqtt + usui + noilong + vitri3,
                  family = binomial, data = e1)

# Stepwise forward
model_step <- step(model_start, 
                   scope = list(lower = model_start, upper = model_full), 
                   direction = "forward")
## Start:  AIC=154.36
## gpb ~ 1
## 
##               Df Deviance    AIC
## + vitri3       1   137.48 141.48
## + noilong      1   138.88 142.88
## + trenp        1   144.10 148.10
## + tn           1   144.64 148.64
## + pqtt         1   144.78 148.78
## + bo4          2   143.28 149.28
## + usui         1   145.55 149.55
## + hang         1   145.91 149.91
## + chanan       1   147.10 151.10
## + notmo        1   147.58 151.58
## + nt           1   147.85 151.85
## + daunguc      1   148.15 152.15
## + trent        1   148.20 152.20
## + xltt         1   148.75 152.75
## + hutthuocla2  2   147.85 153.85
## <none>             152.36 154.36
## 
## Step:  AIC=141.48
## gpb ~ vitri3
## 
##               Df Deviance    AIC
## + trenp        1   129.77 135.77
## + daunguc      1   131.06 137.06
## + pqtt         1   131.59 137.59
## + hang         1   132.06 138.06
## + tn           1   132.20 138.20
## + bo4          2   130.46 138.46
## + chanan       1   133.05 139.05
## + notmo        1   134.00 140.00
## + trent        1   134.11 140.11
## + noilong      1   134.65 140.65
## + xltt         1   135.00 141.00
## + nt           1   135.04 141.04
## <none>             137.48 141.48
## + usui         1   137.12 143.12
## + hutthuocla2  2   135.31 143.31
## 
## Step:  AIC=135.77
## gpb ~ vitri3 + trenp
## 
##               Df Deviance    AIC
## + tn           1   123.62 131.62
## + daunguc      1   124.26 132.26
## + bo4          2   123.11 133.11
## + hang         1   125.26 133.26
## + chanan       1   126.07 134.07
## + pqtt         1   126.74 134.74
## + notmo        1   127.40 135.40
## + xltt         1   127.63 135.63
## <none>             129.77 135.77
## + noilong      1   128.09 136.09
## + nt           1   128.45 136.45
## + hutthuocla2  2   126.85 136.85
## + trent        1   129.06 137.06
## + usui         1   129.65 137.65
## 
## Step:  AIC=131.62
## gpb ~ vitri3 + trenp + tn
## 
##               Df Deviance    AIC
## + hang         1   119.95 129.95
## + daunguc      1   120.31 130.31
## + pqtt         1   120.47 130.47
## + chanan       1   120.50 130.50
## + bo4          2   118.65 130.65
## <none>             123.62 131.62
## + noilong      1   121.74 131.74
## + notmo        1   121.82 131.82
## + trent        1   122.77 132.76
## + nt           1   123.07 133.07
## + xltt         1   123.22 133.22
## + hutthuocla2  2   121.26 133.26
## + usui         1   123.30 133.30
## 
## Step:  AIC=129.95
## gpb ~ vitri3 + trenp + tn + hang
## 
##               Df Deviance    AIC
## + daunguc      1   116.19 128.19
## + bo4          2   114.77 128.77
## + pqtt         1   116.81 128.81
## + chanan       1   116.95 128.96
## + noilong      1   117.53 129.53
## <none>             119.95 129.95
## + notmo        1   118.28 130.28
## + trent        1   118.29 130.29
## + hutthuocla2  2   116.85 130.85
## + xltt         1   119.41 131.41
## + nt           1   119.66 131.66
## + usui         1   119.82 131.82
## 
## Step:  AIC=128.19
## gpb ~ vitri3 + trenp + tn + hang + daunguc
## 
##               Df Deviance    AIC
## + bo4          2   110.90 126.90
## + pqtt         1   113.19 127.19
## + notmo        1   113.64 127.64
## + noilong      1   113.67 127.67
## + chanan       1   113.87 127.87
## <none>             116.19 128.19
## + trent        1   114.68 128.68
## + hutthuocla2  2   113.29 129.29
## + xltt         1   115.59 129.59
## + nt           1   116.00 130.00
## + usui         1   116.14 130.14
## 
## Step:  AIC=126.9
## gpb ~ vitri3 + trenp + tn + hang + daunguc + bo4
## 
##               Df Deviance    AIC
## + notmo        1   107.69 125.69
## + chanan       1   108.23 126.23
## + pqtt         1   108.68 126.68
## <none>             110.90 126.90
## + noilong      1   109.53 127.53
## + trent        1   109.72 127.72
## + hutthuocla2  2   108.06 128.06
## + xltt         1   110.60 128.60
## + nt           1   110.81 128.81
## + usui         1   110.81 128.81
## 
## Step:  AIC=125.69
## gpb ~ vitri3 + trenp + tn + hang + daunguc + bo4 + notmo
## 
##               Df Deviance    AIC
## + chanan       1   105.45 125.45
## <none>             107.69 125.69
## + pqtt         1   105.87 125.87
## + trent        1   106.51 126.51
## + noilong      1   106.78 126.78
## + hutthuocla2  2   104.80 126.80
## + xltt         1   107.33 127.33
## + nt           1   107.34 127.34
## + usui         1   107.62 127.62
## 
## Step:  AIC=125.45
## gpb ~ vitri3 + trenp + tn + hang + daunguc + bo4 + notmo + chanan
## 
##               Df Deviance    AIC
## + pqtt         1   103.34 125.34
## <none>             105.45 125.45
## + trent        1   104.40 126.40
## + noilong      1   104.57 126.57
## + hutthuocla2  2   102.86 126.86
## + xltt         1   105.20 127.20
## + nt           1   105.21 127.21
## + usui         1   105.44 127.44
## 
## Step:  AIC=125.34
## gpb ~ vitri3 + trenp + tn + hang + daunguc + bo4 + notmo + chanan + 
##     pqtt
## 
##               Df Deviance    AIC
## + hutthuocla2  2    99.29 125.29
## <none>             103.34 125.34
## + noilong      1   102.34 126.34
## + xltt         1   102.79 126.79
## + nt           1   102.95 126.95
## + trent        1   103.14 127.14
## + usui         1   103.33 127.33
## 
## Step:  AIC=125.29
## gpb ~ vitri3 + trenp + tn + hang + daunguc + bo4 + notmo + chanan + 
##     pqtt + hutthuocla2
## 
##           Df Deviance    AIC
## <none>         99.290 125.29
## + noilong  1   98.125 126.12
## + trent    1   99.049 127.05
## + xltt     1   99.097 127.10
## + nt       1   99.215 127.22
## + usui     1   99.289 127.29
summary(model_step)
## 
## Call:
## glm(formula = gpb ~ vitri3 + trenp + tn + hang + daunguc + bo4 + 
##     notmo + chanan + pqtt + hutthuocla2, family = binomial, data = e1)
## 
## Coefficients:
##                  Estimate Std. Error z value Pr(>|z|)   
## (Intercept)       -3.8530     1.2963  -2.972  0.00296 **
## vitri3Khối u       1.6128     0.5728   2.816  0.00486 **
## trenp             -1.0681     0.5664  -1.886  0.05935 . 
## tn                 0.6032     0.5816   1.037  0.29974   
## hang              -2.1643     1.1470  -1.887  0.05918 . 
## daunguc            0.8692     0.5297   1.641  0.10080   
## bo41               2.0534     1.2043   1.705  0.08819 . 
## bo42               2.5895     1.1749   2.204  0.02752 * 
## notmo             -1.0061     0.6501  -1.547  0.12174   
## chanan             0.9276     0.6138   1.511  0.13072   
## pqtt               1.1950     0.6479   1.845  0.06511 . 
## hutthuocla2co      0.9271     0.5671   1.635  0.10210   
## hutthuocla2Ngung   1.4688     0.8844   1.661  0.09677 . 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 152.36  on 111  degrees of freedom
## Residual deviance:  99.29  on  99  degrees of freedom
## AIC: 125.29
## 
## Number of Fisher Scoring iterations: 5