without extending or implementing anything from this package. Truly annotation driven. Note: Not implemented yet is clustering algorithms.
extend RecordInterface. Also added MLR implementation that can handle GLS or OLS by setting a flag (and adding a covariance matrix if you're doing GLS).
gathering to a smaller footprint, threadsafe implementation (Commons Math SynchronizedSummaryStatistics, instead of DescriptiveStatistics).
classes and refactored Cluster to extend RecordList since it is a specialized RecordList.
classification problems. Also updated Readme to include an example (hopefully nicely formated).
extending Record class, removed unneeded classes.
double arrays. Also added test regression dataset.
to use as predicted value, ect. Updated RecordSet to force record validation to exclude records with NaN and Infinite Features to help assure successful modeling later.
statitics about the collections and cacheing that information when possible without carrying around a bunch of extra weight.
whole bunch more... classes, tests, improved tests, changes to architecture.
Time Feature ENUM to make configuration more straightforward.