오늘의 ADP 준비 / 2022-08-02 화(D54)

Scheffe 검정

 

Contrasts 개념에 대해서 이해하려고 자료를 좀 더 찾아보았습니다. 충분히 이해하려면 아직도  한참 더 공부해야겠어요.

출처 : https://www.stat.purdue.edu/~ghobbs/STAT_512/Lecture_Notes/ANOVA/Topic_23.pdf

 

기출 풀이

[Adp 실기 기출 풀이] 22회 다항회귀 3차까지 (풀이중...)

https://it-freelancer.tistory.com/83

 

다항회귀분석

국민대 곽기영 교수님 유튜브 영상 공부. abline lines 함수 기능 공부가 조금 더 필요해요.

로에스추세선 처리가 smoother="loessLine" 이 아니라 smoother=loessLine 이란 점 주의해야 합니다.

library(car)
str(Prestige)
Prestige.lm <- lm(income ~ education, data=Prestige)
summary(Prestige.lm)

windows(width=12, height=8)
plot(income ~ education, data=Prestige,
     col="cornflowerblue",pch=19,
     xlab="education (years)",ylab="income (dollars)",
     main ="Education and Income")
abline(Prestige.lm, col = "salmon", lwd = 2)

lm(income ~ education, data=Prestige,
   subset=(education > mean(Prestige$education)))

lm(income ~ education, data=Prestige,
   subset=(education <= mean(Prestige$education)))

scatterplot(income ~ education, data=Prestige,
            pch=19,col="orangered",cex=1.2,
            regLine=list(method=lm,lty=2,lwd=3,col="royalblue"),
            smooth=list(smoother=loessLine,spread=FALSE,
                        lty.smooth=1,lwd.smooth=3,col.smooth="green3"))

Prestige.poly <- lm(income ~ education + I(education^2), data=Prestige)
summary(Prestige.poly)
library(dplyr)
lines(arrange(data.frame(Prestige$education,fitted(Prestige.poly)),
              Prestige$education), col="cornflowerblue", lwd=2)

str(faithful)
scatterplot(eruptions ~ waiting, data=faithful,
            pch=19,col="deepskyblue",cex=1.2,
            regLine=list(method=lm,lty=2,lwd=3,col="royalblue"),
            smooth=list(smoother=loessLine,spread=FALSE,
                        lty.smooth=1,lwd.smooth=3,col.smooth="coral"))

faithful.poly <- lm(eruptions ~ waiting + I(waiting^2) + I(waiting^3), 
                    data=faithful)
summary(faithful.poly)

faithful.lm <- lm(eruptions ~ waiting, 
                    data=faithful)
summary(faithful.lm)

출처: 통계데이터분석 - 회귀분석 - 다항회귀분석 https://youtu.be/53N4NQ1bgCA