The current olsRegress function works on an in-memory matrix. It would be nice to also have an online updating multivariate regression implementation that works with data sets of any size.
This ticket will add support for Miller Updating Regression, which can be used with data sets of any size. An OnlineRegressionStream Stream Decorator will be added to support this functionality. The function name will likely be called onlineRegress.
The implementation will be provided by Apache Commons Math.