Corrected minor performance issue in HClustering. Removed ICluster as an interface (useless here). Added distance metric to KMeans (to remove and usightly forced euclidian calculation).

Fixing class visibility

Adding unit test

Adding hierarchical clustering algo with several linking algos

Sealed distance/metric classes

Added distance/similarity metrics in preparation for making ML algorithms distance/similarity measure agnostic

Converted Test_Continuous_Split into many tests performing one type of assertion.

Added root TestResults Ignore

Fixed bug in test

Fixed failing tests for Serialization

Added more thorough continous DT test. Made changes to vector stacking API in order to make it more intuitive.

Added additional code to solidify DT's in preparation for advanced DT visualizer

Cleaned up KMeans, made some tweaks to Kernel Perceptron and continued improvement of DT's (which are NOT done).

Added vector segmentation to limit fan-out witdh of dt's

Added base learner to train several models and compute accuracy using 80/20 split

Fixed licensing headers

Added (un-tested) kernel perceptron and stopword removal from dictionaries

Added implicit vector conversion from float along with modified tests

Minor updates/changes to DT implementation. Store offset rather than entire Property description in nodes

Added DTPredictor functionality. Added Load/Save DT functionality. Fixed a few DT bugs.

Finished DT implementation on the model side.

Added impurity measures in preparation for decision tree algorithm

added TestResults to .hgignore

Added seeds for DT's. Also added implicit vector and matrix conversions

Travelling salesman problem