Differences

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

--- c:ma:2019:multiple_regression_exercise [2019/11/08 10:51] – [Ex. 2] hkimscil
+++ c:ma:2019:multiple_regression_exercise [2021/11/11 10:14] (current) – [Ex. 2] hkimscil
@@ Line 6: / Line 6: @@
   * Use ''?Carseats'' command for the explanation of the dataset
   * Use ''str'' function to see the characteristic of each variable. Make it sure that ''SelvesLoc'' variable should be factor, not int or anything.
+  * 변인설명을 토대로 가설만들기
+    * 종속변인 = Sales
+    * 독립변인 = 숫자변인 1 + 종류변인 1  (조별 선택)
+    * Multiple regression without interactin
+    * Multiple regression with interaction
+  * 가설 만들기
+    * 종속변인 Sales
+    * 독립변인 여러개 (interaction 없이)
+    * Modeling 해 볼 것
-  * Make a full model (with all variables) then reduce down the model until you find it fitted.
+see [[:hierarchical regression]]
-  * Make a null model (with no variables) then, build up the model with additional IVs until you find a fitted model.
+see also [[:statistical regression methods]] <- 많이 쓰이지 않음
-  * Can we use ''step'' or ''stepAIC'' (MASS package needed) function?
-  * Interpret the result
+  * <del>Make a full model (with all variables) then reduce down the model until you find it fitted.</del>
+  * <del>Make a null model (with no variables) then, build up the model with additional IVs until you find a fitted model.</del>
+  * <del>Can we use ''step'' or ''stepAIC'' (MASS package needed) function?</del>
+  * <del>Interpret the result</del>
+<del>> step(lm.full, direction="back")</del>
-> step(lm.full, direction="back")
+[[./eg_script]]
 ===== Ex. 2 =====
@@ Line 18: / Line 32: @@
   * load the tidyverse
   * ''%%install.packages("car")%%''
-  * ''%%data("Salaries", package = "car")%%''
+  * ''%%install.packages("carData")%%''
-  * Use a dataset Salaries
+  * load the car and the carData
-  * describe the data set
+  * ''%%data("Salaries", package = "carData")%%''
+  * ''%%?Salaries%%''
+    * explain what it is
+    * describe the data set
 ----
   * Regress sex variable on salary variable
   * Write the regression model
-  * Discuss the difference
+  * Discuss the difference male and female (sex)
   * Use rank variable for the same purpose
-  * --
+  * Write the regression model
-  * Use yrs.service + rank + discipline + sex
+  * Regress rank + sex on salary
-  * on salary
   * How do you interpret the result?
+  * And regress rank + sex + rank:sex on salary
+  * How do you interpret this result?
+  * Do factorial ANOVA test with rank and sex on salary
+  * How do you interpret the result?
+  * Test regression model of your own choice
+  * Interpret the result
 -----
 위의 Salaries 데이터사용이 안 될 때
@@ Line 41: / Line 65: @@
 Please copy and paste the proper r command and output to a txt file (use notepad or some other text editing program). You could use MS Word, but, please make it sure that you use type-setting fonts such as "Courier New." The below output, as an example, includes the r command ''%%head(Salaries)%%'' and the output.
 <code>
-> head(Salaries)
-       rank discipline yrs.since.phd yrs.service  sex salary
-      Prof          B            19          18 Male 139750
-      Prof          B            20          16 Male 173200
-  AsstProf          B             4           3 Male  79750
-      Prof          B            45          39 Male 115000
-      Prof          B            40          41 Male 141500
-AssocProf          B             6           6 Male  97000
-</code>
-<code>> lm.sal.sex <- lm(salary ~ sex, data=Salaries)
-> summary(lm.sal.sex)
-Call:
-lm(formula = salary ~ sex, data = Salaries)
-Residuals:
-   Min     1Q Median     3Q    Max
--57290 -23502  -6828  19710 116455
-Coefficients:
-            Estimate Std. Error
-(Intercept)   101002       4809
-sexMale        14088       5065
-            t value Pr(>|t|)
-(Intercept)  21.001  < 2e-16 ***
-sexMale       2.782  0.00567 **
----
-Signif. codes:
-‘***’ 0.001 ‘**’ 0.01 ‘*’
-.05 ‘.’ 0.1 ‘ ’ 1
-Residual standard error: 30030 on 395 degrees of freedom
-Multiple R-squared:  0.01921,	Adjusted R-squared:  0.01673
-F-statistic: 7.738 on 1 and 395 DF,  p-value: 0.005667
-</code>
-<code>> lm.sal.rank <- lm(salary ~ rank, data=Salaries)
-> summary(lm.sal.rank)
-Call:
-lm(formula = salary ~ rank, data = Salaries)
-Residuals:
-   Min     1Q Median     3Q    Max
--68972 -16376  -1580  11755 104773
-Coefficients:
-              Estimate Std. Error
-(Intercept)      80776       2887
-rankAssocProf    13100       4131
-rankProf         45996       3230
-              t value Pr(>|t|)
-(Intercept)    27.976  < 2e-16 ***
-rankAssocProf   3.171  0.00164 **
-rankProf       14.238  < 2e-16 ***
----
-Signif. codes:
-‘***’ 0.001 ‘**’ 0.01 ‘*’
-.05 ‘.’ 0.1 ‘ ’ 1
-Residual standard error: 23630 on 394 degrees of freedom
-Multiple R-squared:  0.3943,	Adjusted R-squared:  0.3912
-F-statistic: 128.2 on 2 and 394 DF,  p-value: < 2.2e-16
-> </code>
-<code>> summary(lm.sal.many)
-Call:
-lm(formula = salary ~ yrs.service + rank + discipline + sex,
-    data = Salaries)
-Residuals:
-   Min     1Q Median     3Q    Max
--64202 -14255  -1533  10571  99163
-Coefficients:
-              Estimate Std. Error t value Pr(>|t|)
-(Intercept)   68351.67    4482.20  15.250  < 2e-16 ***
-yrs.service     -88.78     111.64  -0.795 0.426958
-rankAssocProf 14560.40    4098.32   3.553 0.000428 ***
-rankProf      49159.64    3834.49  12.820  < 2e-16 ***
-disciplineB   13473.38    2315.50   5.819 1.24e-08 ***
-sexMale        4771.25    3878.00   1.230 0.219311
----
-Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
-Residual standard error: 22650 on 391 degrees of freedom
-Multiple R-squared:  0.4478,	Adjusted R-squared:  0.4407
-F-statistic: 63.41 on 5 and 391 DF,  p-value: < 2.2e-16
-> </code>
 ====== Discussion ======
 Common topics