Differences between revisions 1 and 3 (spanning 2 versions)
⇤ ← Revision 1 as of 2017-02-09 20:03:09 →
Size: 776
Editor: DavidOwen
Comment: Added notes
← Revision 3 as of 2017-02-09 20:04:59 → ⇥
Size: 792
Editor: DavidOwen
Comment:
Deletions are marked like this. Additions are marked like this.
Line 7: Line 7:
{{{
Line 21: Line 22:
}}}
Line 24: Line 25:
{{{
Line 25: Line 27:
}}}

Feb 09, 2017

/Attendees

Review of Chapter 3

Played with the advertising data. 

d <- read.csv("Advertising.csv")
plot(Sales ~ TV, d)
m <- lm(Sales ~ TV, d)
summary(m)
abline(m)
matlines(predict(m, newdata=data.frame(TV=0:300), interval="p"))
d$x <- log(d$TV)
d$y <- log(d$Sales)
m <- lm(y ~ x, d)
summary(m)
plot(y ~ x, d)
matlines(log(0:300), predict(m, newdata=data.frame(x=log(0:300)), interval="p"))
plot(Sales ~ TV, d)
matlines(0:300, exp(predict(m, newdata=data.frame(x=log(0:300)), interval="p")))

Talked about diagnosing model problems. 

plot(m)

But also, 4-plots and 6-plots as seen in the NIST handbook.

Next Time

MeetingNotes/2017-02 (last edited 2017-02-09 20:04:59 by DavidOwen)