Differences

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

--- beta_coefficients [2019/05/21 11:43] – hkimscil
+++ beta_coefficients [2020/12/09 18:47] (current) – [e.g.] hkimscil
@@ Line 1: / Line 1: @@
-====== Beta coefficients in regression ======
+====== Beta coefficients in linear regression ======
-$$ \beta = b * \frac{sd(x)}{sd(y)} $$
+{{:pasted:20190521-113150.png?200}}
+\begin{align*}
+\large{\beta = b * \frac{sd(x)}{sd(y)}} \
+\end{align*}
 <code>
@@ Line 10: / Line 15: @@
 </code>
-<code>lm.gpa.clep <- lm(gpa ~ clep, data = tests)
+<code>lm.gpa.clepsat <- lm(gpa ~ clep + sat, data = tests)
-summary(lm.gpa.clep)
+summary(lm.gpa.clepsat)
-</code>
-<code>
 Call:
-lm(formula = gpa ~ clep, data = tests)
+lm(formula = gpa ~ clep + sat, data = tests)
 Residuals:
       Min        1Q    Median        3Q       Max
--0.190496 -0.141167 -0.002376  0.110847  0.225207
+-0.197888 -0.128974 -0.000528  0.131170  0.226404
 Coefficients:
-            Estimate Std. Error t value Pr(>|t|)
+              Estimate Std. Error t value Pr(>|t|)
-(Intercept)  1.17438    0.38946   3.015 0.016676 *
+(Intercept)  1.1607560  0.4081117   2.844   0.0249 *
-clep         0.06054    0.01177   5.144 0.000881 ***
+clep         0.0729294  0.0253799   2.874   0.0239 *
+sat         -0.0007015  0.0012564  -0.558   0.5940
 ---
 Signif. codes:
 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
-Residual standard error: 0.1637 on 8 degrees of freedom
+Residual standard error: 0.1713 on 7 degrees of freedom
-Multiple R-squared:  0.7679,	Adjusted R-squared:  0.7388
+Multiple R-squared:  0.7778,	Adjusted R-squared:  0.7143
-F-statistic: 26.46 on 1 and 8 DF,  p-value: 0.0008808
+F-statistic: 12.25 on 2 and 7 DF,  p-value: 0.005175
+>
 </code>
+<code>> sd.clep <- sd(clep)
+> sd.sat <- sd(sat)
+> sd.gpa <- sd(gpa)
+> lm.gpa.clepsat <- lm(gpa ~ clep + sat, data = tests)
+> summary(lm.gpa.clepsat)
+Call:
+lm(formula = gpa ~ clep + sat, data = tests)
+Residuals:
+      Min        1Q    Median        3Q       Max
+-0.197888 -0.128974 -0.000528  0.131170  0.226404
+Coefficients:
+              Estimate Std. Error t value Pr(>|t|)
+(Intercept)  1.1607560  0.4081117   2.844   0.0249 *
+clep         0.0729294  0.0253799   2.874   0.0239 *
+sat         -0.0007015  0.0012564  -0.558   0.5940
+---
+Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
+Residual standard error: 0.1713 on 7 degrees of freedom
+Multiple R-squared:  0.7778,	Adjusted R-squared:  0.7143
+F-statistic: 12.25 on 2 and 7 DF,  p-value: 0.005175
+> b.clep <- 0.0729294
+> b.sat <- -0.0007015
+> beta.clep <- b.clep * (sd.clep/sd.gpa)
+> beta.sat <- b.sat * (sd.sat/sd.gpa)
+> lm.beta(lm.gpa.clepsat)
+Call:
+lm(formula = gpa ~ clep + sat, data = tests)
+Standardized Coefficients::
+(Intercept)        clep         sat
+.0000000   1.0556486  -0.2051189
+> beta.clep
+[1] 1.055648
+> beta.sat
+[1] -0.2051187
+>
+</code>
+====== e.g. ======
+<code>
+# get marketing data
+marketing <- read.csv("http://commres.net/wiki/_media/marketing_from_datarium.csv")
+head(marketing)
+# note that I need - X to get rid of X column in the marketing data
+mod <- lm(sales ~ . - X, data=marketing)
+summary(mod)
+</code>
+<code>
+> marketing <- read.csv("http://commres.net/wiki/_media/marketing_from_datarium.csv")
+> head(marketing)
+  X youtube facebook newspaper sales
+1  276.12    45.36     83.04 26.52
+2   53.40    47.16     54.12 12.48
+3   20.64    55.08     83.16 11.16
+4  181.80    49.56     70.20 22.20
+5  216.96    12.96     70.08 15.48
+6   10.44    58.68     90.00  8.64
+# note that I need - X to get rid of X column in the marketing data
+> mod <- lm(sales ~ . - X, data=marketing)
+> summary(mod)
+Call:
+lm(formula = sales ~ . - X, data = marketing)
+Residuals:
+     Min       1Q   Median       3Q      Max
+-10.5932  -1.0690   0.2902   1.4272   3.3951
+Coefficients:
+             Estimate Std. Error t value Pr(>|t|)
+(Intercept)  3.526667   0.374290   9.422   <2e-16 ***
+youtube      0.045765   0.001395  32.809   <2e-16 ***
+facebook     0.188530   0.008611  21.893   <2e-16 ***
+newspaper   -0.001037   0.005871  -0.177     0.86
+---
+Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
+Residual standard error: 2.023 on 196 degrees of freedom
+Multiple R-squared:  0.8972,	Adjusted R-squared:  0.8956
+F-statistic: 570.3 on 3 and 196 DF,  p-value: < 2.2e-16
+</code>
+<code>
+install.packages(lm.beta)
+library(lm.beta)
+lm.beta(mod)
+</code>
+<code>
+lm.beta(mod)
+Call:
+lm(formula = sales ~ . - X, data = marketing)
+Standardized Coefficients::
+ (Intercept)      youtube     facebook    newspaper
+.000000000  0.753065912  0.536481550 -0.004330686
+>
+</code>
+These beta coefficients also can be got from the coefficents from standardized data.
+<code>
+mod.formula <- sales ~ youtube + facebook + newspaper
+all.vars(mod.formula)
+marketing.temp <- sapply(marketing[ , all.vars(mod.formula)], scale)
+head(marketing.temp)
+mod.scaled <- lm(sales ~ ., data=marketing.scaled)
+head(marketing.scaled)
+coefficients(mod.scaled)
+</code>
+<code>> mod.formula <- sales ~ youtube + facebook + newspaper
+> all.vars(mod.formula)
+[1] "sales"     "youtube"   "facebook"  "newspaper"
+> marketing.temp <- sapply(marketing[ , all.vars(mod.formula)], scale)
+> head(marketing.temp)
+          sales     youtube   facebook newspaper
+[1,]  1.5481681  0.96742460  0.9790656 1.7744925
+[2,] -0.6943038 -1.19437904  1.0800974 0.6679027
+[3,] -0.9051345 -1.51235985  1.5246374 1.7790842
+[4,]  0.8581768  0.05191939  1.2148065 1.2831850
+[5,] -0.2151431  0.39319551 -0.8395070 1.2785934
+[6,] -1.3076295 -1.61136487  1.7267010 2.0408088
+> mod.scaled <- lm(sales ~ ., data=marketing.scaled)
+> head(marketing.scaled)
+       sales     youtube   facebook newspaper
+  1.5481681  0.96742460  0.9790656 1.7744925
+-0.6943038 -1.19437904  1.0800974 0.6679027
+-0.9051345 -1.51235985  1.5246374 1.7790842
+  0.8581768  0.05191939  1.2148065 1.2831850
+-0.2151431  0.39319551 -0.8395070 1.2785934
+-1.3076295 -1.61136487  1.7267010 2.0408088
+> coefficients(mod.scaled)
+  (Intercept)       youtube      facebook     newspaper
+-5.034110e-16  7.530659e-01  5.364815e-01 -4.330686e-03
+>
+> </code>
+check out that
+''lm.beta(mod) == coefficients(mod.scaled)''
+and
+베타를 구하고 나면 서로의 계수값을 절대비교할 수 있다.
+''lm.beta(mod)''의 아웃풋을 보고
+''youtube'', ''facebook'', and ''newspaper'' 순으로 설명력을 갖는다고 말할 수 있다 (혹은 newspaper를 분석에서 제외하고 다시 분석하여 둘만의 설명력을 보는 것도 방법이다.