https://lms.onnocenter.or.id/wiki/index.php?action=history&feed=atom&title=R_Regression%3A_stepwise_regression_essentials
R Regression: stepwise regression essentials - Revision history
2026-04-21T02:56:33Z
Revision history for this page on the wiki
MediaWiki 1.45.1
https://lms.onnocenter.or.id/wiki/index.php?title=R_Regression:_stepwise_regression_essentials&diff=57699&oldid=prev
Onnowpurbo: Created page with " # Ref: http://www.sthda.com/english/articles/37-model-selection-essentials-in-r/154-stepwise-regression-essentials-in-r/ library(tidyverse) library(caret) library(leaps)..."
2019-12-02T01:46:22Z
<p>Created page with " # Ref: http://www.sthda.com/english/articles/37-model-selection-essentials-in-r/154-stepwise-regression-essentials-in-r/ library(tidyverse) library(caret) library(leaps)..."</p>
<p><b>New page</b></p><div> # Ref: http://www.sthda.com/english/articles/37-model-selection-essentials-in-r/154-stepwise-regression-essentials-in-r/<br />
<br />
library(tidyverse)<br />
library(caret)<br />
library(leaps)<br />
<br />
<br />
<br />
# Computing stepwise regression<br />
library(MASS)<br />
# Fit the full model <br />
full.model <- lm(Fertility ~., data = swiss)<br />
# Stepwise regression model<br />
step.model <- stepAIC(full.model, direction = "both", <br />
trace = FALSE)<br />
summary(step.model)<br />
<br />
models <- regsubsets(Fertility~., data = swiss, nvmax = 5,<br />
method = "seqrep")<br />
summary(models)<br />
<br />
<br />
<br />
<br />
# Set seed for reproducibility<br />
set.seed(123)<br />
# Set up repeated k-fold cross-validation<br />
train.control <- trainControl(method = "cv", number = 10)<br />
# Train the model<br />
step.model <- train(Fertility ~., data = swiss,<br />
method = "leapBackward", <br />
tuneGrid = data.frame(nvmax = 1:5),<br />
trControl = train.control<br />
)<br />
step.model$results<br />
<br />
# best tuning values (nvmax)<br />
step.model$bestTune<br />
<br />
summary(step.model$finalModel)<br />
<br />
<br />
# regression coefficients of the final model (id = 4)<br />
coef(step.model$finalModel, 4)<br />
lm(Fertility ~ Agriculture + Education + Catholic + Infant.Mortality, <br />
data = swiss)<br />
<br />
<br />
<br />
<br />
<br />
# ALTERNATIVE<br />
library(MASS)<br />
res.lm <- lm(Fertility ~., data = swiss)<br />
step <- stepAIC(res.lm, direction = "both", trace = FALSE)<br />
step<br />
<br />
# Train the model<br />
step.model <- train(Fertility ~., data = swiss,<br />
method = "lmStepAIC", <br />
trControl = train.control,<br />
trace = FALSE<br />
)<br />
# Model accuracy<br />
step.model$results<br />
# Final model coefficients<br />
step.model$finalModel<br />
# Summary of the model<br />
summary(step.model$finalModel)<br />
<br />
<br />
<br />
==Referensi==<br />
<br />
* http://www.sthda.com/english/articles/37-model-selection-essentials-in-r/154-stepwise-regression-essentials-in-r/<br />
<br />
==Pranala Menarik==<br />
<br />
* [[R]]</div>
Onnowpurbo