Re: [Wekalist] (no subject)

[Guillaume M. <gm...@pr...>]

Hi,

The Weka developers have done a great job in making as much abstraction as possible in Weka, so the steps involved to apply a regression algorithm is almost identical as the ones to apply a classification algorithm. Just i) use a dataset with a numeric class, ii) select a regression algo iii) eventually learn how to tune the parameters of this specific algo (requires to know deeply how it works).

If you want to learn how to use Weka, you should *really* consider at least watching the videos from the Weka MOOC: https://www.youtube.com/user/WekaMOOC , or register to the course. It's really well done, with simple explanations and covers a sufficiently wide kinds of algorithms/problems that you should be almost autonomous in the remaining of your learning. There's also plenty of documentation in the WekaManual.pdf (which comes with every installation).

Cheers,

GM
PS: before sending files to a mailing list, remember that 1) hundreds of thousands of people will receive them (whereas only a few will actually be interested in them), some people do not have high bandwidth or email storage capacities, so even sending middle-sized files (worse: blobs like .xlsx) you might overload their network/email account.