Help me, Statistical LJ

[drwex]

I know I should know the answer to this, but I'm failstats today...

Given that I have two series of numbers (a1, b1) (a2, b2) ... (aN, bN) I want to know the degree of correlation or the predictive value of aN.

Comments

[gsh]

is http://en.wikipedia.org/wiki/Correlation_and_dependence what you are looking for?

Sort of

[drwex]

as usual it's too much information. I need a sentence or two, not an encyclopedia entry defining the topic.

[sunstealer.livejournal.com]

If I understand you- you're looking for the predictive power of variable one on variable 2, i.e. how much of the variance in bN is related to aN?

If so, that's a linear regression. Most spreadsheet applications will do that.

Thank you

[drwex]

I think that's indeed what I'm looking for. The collection is slow since I get one a and one b per day but eventually I'll have enough data to throw together a rough line.

[sunstealer.livejournal.com]

and when you run a regression, you'll get something called an r value (your correlation coefficient), the higher that is, the higher the predictive value of your independent variable on your dependent.

Re: Thank you

[sunstealer.livejournal.com]

sweet. good luck.

[sykotropic.livejournal.com]

I second what sunstealer said. Go for it. When you can.