CURM Fourth Meeting, 9/23/2008
Jump to navigation
Jump to search
Agenda
- Last time: we had a look at a couple of graphics in Jon Hastings's power point (including one that showed a regression of wing length on body mass). We talked about the "banana" in the shape of the data that suggested the linear model was inappropriate.
- Today, let's look at basic regression, and how it works. We'll also talk about some of the usual transformations, which are based on our understanding of the "function zoo".
- To examine regression, we'll use a data set from the QELP website. Our sample data set focuses on the question How do linear dimensions of an animal correlate with mass?
- To use StatCrunch with the data, you need to
- click the link for StatCrunch, and then
- use the passcode "heyu";
- then you can load the data using this link to the data: http://www.seattlecentral.edu/qelp/sets/023/s023web.txt (any other web-based data set also works, so Statcrunch is a nice option for looking at data on the web).
- We'll start with linear regression,
- Make one transformation, then
- Make another!
- A comment on one relevant variant of linear regression, from the Boyce project
- To use StatCrunch with the data, you need to
- What issues have arisen so far for you? Anything worth mentioning now?
- Grayson?
- Katie?
- Tim?
- Any progress on the two outstanding Problems of the Week?
- Any luck with the randomization/numerics on Problem 1?
- We talked about generalizing Grayson's idea for Problem 2 last time. Any luck?
- Aimee Krug liked the "points" Problem 2 so much he posted his solution.
- Here's a new one! Let's spend a little time with it....