Quote:
Originally Posted by kramarczyk
For the record I am not suggesting that excel's linear fit should be used as I have no idea what data is being worked with. The question was about finding correlation between data ranges in a spreadsheet. I was simply pointing out that excel has built in functions for that... I see no reason to reinvent the wheel if one is just trying to convert the data into information. However, if one is trying to learn about the technique then, yeah, grab a book.
I also found this page that gives the excel calcs for other trendlines... http://j-walk.com/ss/excel/tips/tip101.htm |
Yeeesss... But, you see, everything Excel does is still basic single variable linear regression. You'll notice the formulas you linked to there are just a series of linear regressions on transformed data. If you need to do a logistic regression, you're flat out of luck. And if reality fundamentally depends on two independent variables, then you're unlikely to get any useful information out of Excel. Since variations in your data due to the second variable will mostly appear as so much noise if your regression only accounts for one variable. Unless you only look at data points where all other variables are the same, which isn't terribly useful. On the whole, you're much more likely to get useful information out of the data by actually, you know, doing statistics on it.