Scribd
Upload
6K views

Regression Splines

This document discusses different regression spline techniques for modeling the relationship between wage and age using data from the Wage dataset. It compares cubic splines with pre-specified knots, natural splines with knots at quantiles, smoothing splines, and local regression. For each technique, R code is provided to fit the model and visualize the results, demonstrating how the different spline approaches can capture the nonlinear pattern between wage and age.

Uploaded by

api-285777244
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
6K views

Regression Splines

This document discusses different regression spline techniques for modeling the relationship between wage and age using data from the Wage dataset. It compares cubic splines with pre-specified knots, natural splines with knots at quantiles, smoothing splines, and local regression. For each technique, R code is provided to fit the model and visualize the results, demonstrating how the different spline approaches can capture the nonlinear pattern between wage and age.

Uploaded by

api-285777244
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Regression Splines

YIK LUN, KEI


[email protected]
This paper is a lab from the book called An Introduction to Statistical Learning
with Applications in R. All R codes and comments below are belonged to the
book and authors.

Cubic splines are produced by default

200
100
50

wage

300

library(ISLR)
attach(Wage)
library(splines)
agelims=range(age)
age.grid=seq(agelims[1],agelims[2])
fit=lm(wage~bs(age, knots =c(25 ,40 ,60)), data=Wage) #prespecified knots
pred=predict(fit ,newdata =list(age =age.grid),se=T)
plot(age,wage ,col =" gray ")
lines(age.grid ,pred$fit,col="blue" ,lwd =2)
lines(age.grid ,pred$fit +2* pred$se, lty ="dashed")
lines(age.grid ,pred$fit -2* pred$se, lty ="dashed")

20

30

40

50
age

60

70

80

Natural spline with knots at uniform quantiles


fit2=lm(wage~ns(age ,df =4) ,data=Wage)
attr(ns(age,df=4) ,"knots")
##
25%
50%
75%
## 33.75 42.00 51.00

200
50

100

wage

300

pred2=predict(fit2,newdata =list(age=age.grid),se=T)
plot(age,wage ,col =" gray ")
lines(age.grid ,pred$fit,col="blue" ,lwd =2)
lines(age.grid, pred2$fit ,col ="red",lwd =2)

20

30

40

50

60

70

age

Smoothing Spline
fit=smooth.spline(age ,wage ,df =16)
fit2=smooth.spline(age ,wage, cv=T)
## Warning in smooth.spline(age, wage, cv = T): cross-validation with non## unique 'x' values seems doubtful
2

80

fit2$df
## [1] 6.794596
plot(age,wage,xlim=agelims,cex =.5, col ="darkgrey",main="Smoothing Spline")
lines(fit ,col ="red ",lwd =2)
lines(fit2 ,col =" blue",lwd =2)
legend ("topright",legend =c("16 DF" ,"6.8 DF"),col=c("red","blue"),lty =1, lwd =2, cex =.8)

Smoothing Spline

200
50

100

wage

300

16 DF
6.8 DF

20

30

40

50

60

70

80

age

Local Regression
fit=loess (wage~age ,span =.2, data=Wage)
fit2=loess(wage~age ,span =.5, data=Wage)
plot(age,wage ,xlim=agelims ,cex =.5, col ="darkgrey",main="Local Regression")
lines(age.grid,predict(fit,data.frame(age=age.grid)),col ="red",lwd =2)
lines(age.grid,predict(fit2 ,data.frame(age=age.grid)),col ="blue",lwd =2)
legend ("topright",legend =c("Span = 0.2","Span = 0.5") ,col=c("red","blue"),lty =1, lwd =2, cex =.8)

Local Regression

200
50

100

wage

300

Span = 0.2
Span = 0.5

20

30

40

50

60

70

80

age

Reference:
James, Gareth, et al. An introduction to statistical learning. New
York: springer, 2013.

You might also like