It is close to this method: http://alias-i.com/lingpipe/demos/tutorial/logistic-regression/read-me.html
The difference is that learning rates for individual coefficients are adapted online. Can you say a bit more about how much data you are looking at? What kind of data? The Mahout logistic regression is optimized for very large numbers of sparse inputs. It works on other data, but doesn't have a huge advantage there. On Sun, May 29, 2011 at 4:38 PM, skasab2s <[email protected]>wrote: > Hello, > > I am currently testing Apache Mahout for usage of its Logistic Regression > implementation in a medical application. I have the following question: > > Which formula does Mahout use in order to calculate the Logistic Regression > Coefficients when learning from data ? Can you give me some more literature > references about how the learning process happens ? > > Thanks a lot and best regards, > > Svetlomir Kasabov. > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Logistic-Regression-Coeffcients-Calculation-tp3000343p3000343.html > Sent from the Mahout User List mailing list archive at Nabble.com. >
