Discussion in 'Strategy Development' started by Michael Black, Apr 5, 2010.
Has anyone had experience with MARS? Salford, R, Statistica or others?
Yes there are a few nice packages in this sector? What exactly are you trying to do and what do you need help with?
You should probably have a cocktail or two with MAESTRO.
I have learned that making generalizations from either linear or polynomial regression has not always been always been a good way to find meaningful characterizations of the time series data I have been interested in. In fact, I have learned a great deal dividing and exploring sections of the independent X variable axis rather then making general assumptions from R squared values and T scores from the entire population.
When I was introduced to the idea of MARS at a data mining conference that I recently attended, it seemed at first that this process of exploring data matched the rather crude methods I had been attempting to do by exploring scatter plots manually.
I'm not sure if I entirely understand it, but am curious to learn more.
In R there is a package called earth that does mars...
Separate names with a comma.