Differences

This shows you the differences between two versions of the page.

--- multicolinearity [2018/12/26 02:23] – hkimscil
+++ multicolinearity [2018/12/26 02:49] (current) – [regression test with factors] hkimscil
@@ Line 1: / Line 1: @@
-====== Multi-colinearity ======
+====== Multi-colinearity check in r ======
+required library:
+  * corrplot
+  * mctest
+    * omcdiag
+    * imcdiag
 <code>
 > cps <- read.csv("http://commres.net/wiki/_media/cps_85_wages.csv", header = T, sep = "\t")
@@ Line 70: / Line 77: @@
 > library(corrplot)
 > cps.cor = cor(cps)
-> corrplot.mixed(cps.cor, lower.col = "black", number.cex = .7)
+> corrplot.mixed(cps.cor, lower.col = "black")
 </code>
-{{cps.corplot.png}}
+{{cps.corrplot.png?500}}
 <code>
@@ Line 210: / Line 217: @@
 > </code>
+====== regression test with factors ======
 <code>
+> cps$sex <- factor(cps$sex)
+> cps$union <- factor(cps$union)
+> cps$race <- factor(cps$race)
+> cps$sector <- factor(cps$sector)
+> cps$occupation <- factor(cps$occupation)
+> cps$marr <- factor(cps$marr)
+> str(cps)
+'data.frame':	534 obs. of  11 variables:
+ $ education : int  8 9 12 12 12 13 10 12 16 12 ...
+ $ south     : int  0 0 0 0 0 0 1 0 0 0 ...
+ $ sex       : Factor w/ 2 levels "0","1": 2 2 1 1 1 1 1 1 1 1 ...
+ $ experience: int  21 42 1 4 17 9 27 9 11 9 ...
+ $ union     : Factor w/ 2 levels "0","1": 1 1 1 1 1 2 1 1 1 1 ...
+ $ wage      : num  5.1 4.95 6.67 4 7.5 ...
+ $ age       : int  35 57 19 22 35 28 43 27 33 27 ...
+ $ race      : Factor w/ 3 levels "1","2","3": 2 3 3 3 3 3 3 3 3 3 ...
+ $ occupation: Factor w/ 6 levels "1","2","3","4",..: 6 6 6 6 6 6 6 6 6 6 ...
+ $ sector    : Factor w/ 3 levels "0","1","2": 2 2 2 1 1 1 1 1 2 1 ...
+ $ marr      : Factor w/ 2 levels "0","1": 2 2 1 1 2 1 1 1 2 1 ...
 </code>
 <code>
+> lm4 = lm(log(cps$wage) ~ . -age, data = cps)
+> summary(lm4)
+Call:
+lm(formula = log(cps$wage) ~ . - age, data = cps)
+Residuals:
+     Min       1Q   Median       3Q      Max
+-2.36103 -0.28080  0.00362  0.27793  1.79594
+Coefficients:
+             Estimate Std. Error t value Pr(>|t|)
+(Intercept)  1.194821   0.181804   6.572 1.21e-10 ***
+education    0.066603   0.010060   6.621 8.96e-11 ***
+south       -0.093384   0.041931  -2.227  0.02637 *
+sex1        -0.216934   0.041844  -5.184 3.11e-07 ***
+experience   0.009371   0.001725   5.431 8.63e-08 ***
+union1       0.211506   0.051218   4.129 4.24e-05 ***
+race2       -0.033928   0.099051  -0.343  0.73209
+race3        0.079851   0.057392   1.391  0.16472
+occupation2 -0.364444   0.091500  -3.983 7.78e-05 ***
+occupation3 -0.210295   0.076175  -2.761  0.00597 **
+occupation4 -0.383882   0.080990  -4.740 2.77e-06 ***
+occupation5 -0.050664   0.072717  -0.697  0.48628
+occupation6 -0.265348   0.079969  -3.318  0.00097 ***
+sector1      0.114857   0.054862   2.094  0.03678 *
+sector2      0.093138   0.096514   0.965  0.33499
+marr1        0.062211   0.041025   1.516  0.13002
+---
+Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
+Residual standard error: 0.4278 on 518 degrees of freedom
+Multiple R-squared:  0.3614,	Adjusted R-squared:  0.3429
+F-statistic: 19.54 on 15 and 518 DF,  p-value: < 2.2e-16
+>
 </code>
-<code>
+<code>> lm5 = lm(log(cps$wage) ~ . -age -race, data = cps)
+> summary(lm5)
+Call:
+lm(formula = log(cps$wage) ~ . - age - race, data = cps)
+Residuals:
+     Min       1Q   Median       3Q      Max
+-2.34366 -0.28169 -0.00017  0.29179  1.81158
+Coefficients:
+             Estimate Std. Error t value Pr(>|t|)
+(Intercept)  1.224289   0.172070   7.115 3.73e-12 ***
+education    0.068838   0.009912   6.945 1.14e-11 ***
+south       -0.102588   0.041668  -2.462 0.014139 *
+sex1        -0.213602   0.041842  -5.105 4.65e-07 ***
+experience   0.009494   0.001723   5.510 5.65e-08 ***
+union1       0.202720   0.051009   3.974 8.06e-05 ***
+occupation2 -0.355381   0.091448  -3.886 0.000115 ***
+occupation3 -0.209820   0.076149  -2.755 0.006068 **
+occupation4 -0.385680   0.080855  -4.770 2.40e-06 ***
+occupation5 -0.047694   0.072746  -0.656 0.512351
+occupation6 -0.254277   0.079781  -3.187 0.001523 **
+sector1      0.111458   0.054845   2.032 0.042636 *
+sector2      0.099777   0.096481   1.034 0.301541
+marr1        0.065464   0.041036   1.595 0.111257
+---
+Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
+Residual standard error: 0.4283 on 520 degrees of freedom
+Multiple R-squared:  0.3573,	Adjusted R-squared:  0.3412
+F-statistic: 22.24 on 13 and 520 DF,  p-value: < 2.2e-16
+>
 </code>
+<code>> lm6 = lm(log(cps$wage) ~ . -age -race -occupation -marr -sector, data = cps)
+> summary(lm6)
+Call:
+lm(formula = log(cps$wage) ~ . - age - race - occupation - marr -
+    sector, data = cps)
+Residuals:
+     Min       1Q   Median       3Q      Max
+-2.13809 -0.28681 -0.00078  0.29376  1.96678
+Coefficients:
+             Estimate Std. Error t value Pr(>|t|)
+(Intercept)  0.731792   0.122217   5.988 3.94e-09 ***
+education    0.094096   0.007942  11.848  < 2e-16 ***
+south       -0.111761   0.042857  -2.608 0.009372 **
+sex1        -0.231978   0.039202  -5.918 5.88e-09 ***
+experience   0.011548   0.001680   6.875 1.75e-11 ***
+union1       0.198360   0.051243   3.871 0.000122 ***
+---
+Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
+Residual standard error: 0.4433 on 528 degrees of freedom
+Multiple R-squared:  0.3011,	Adjusted R-squared:  0.2944
+F-statistic: 45.49 on 5 and 528 DF,  p-value: < 2.2e-16
+> </code>