2 Mar 2008 00:22
Re: Sine function won't resolve correct treadline
Ivan Johansen <graph <at> padowan.dk>
2008-03-01 23:22:24 GMT
2008-03-01 23:22:24 GMT
Darren Marshall wrote: > I came across a little problem. I have a data list, the time of sunrise on the > first of the month for twelve months. A perfect sine, right? Well, after > plotting the data on Graph and asking it to make a sinusodial treadline, it > came up with something weird... Here's the data. See if you can replicate it. Graph uses an initial guess to search for the nearest optimum for the constants that gives the best trendline. The problem is that some models, like this one, have more than one local optimum, and Graph will usually find the one closest to the initial guess. Therefore you may need to change the initial guess, which defaults to 1. I was not aware the TI-83 could find trendlines, but I can see that it only works for some specific functions where it apparently knows what to use for the initial guess. I will consider to do something similar in Graph, and add some special functions, like the sinusoidal, and have Graph automatically find an initial guess for them. That would probably make it more user friendly. Best regards Ivan Johansen ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
RSS Feed