MLPY is a high-performance Python/NumPy based package for machine learning.
What's New in This Release: [ read full changelog ]
· New function span_pd(). three_points_pd() deprecated.
New Dtw class (dtw() has been removed):
· Naive and Derivative DTW
· Symmetric, Asymmetric, Quasi-Symmetric implementation with Slope Constraint Condition P=0
· Sakoe-Chiba window condition option
· Linear space-complexity implementation option
· (0, 0) boundary condition option
· canberra() - canberraq(): new option 'dist' returns partial distances
· canberra - canberraq: partial distances to file(s) added
· Documentation improved
· Derivative DTW algorithm fixed
· knn_imputing() inf**2 bug fixed