You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Three big feature updates since last stable version:
The metric is now passed to DistanceMetricClassifier during the prediction.
Added new in-built distance metrics in #16 - now a total of 43 distance metrics are built into DistClassiPy. Additionally, users can still define and use custom distance metrics as needed.
Added a new classifier, the EnsembleDistanceClassifier, which takes in a feature to partition data into quantiles, and then prepares an ensemble of DistanceMetricClassifiers for each quantile based on the best distance metric for that quantile.
Other minor updates include:
Added tests for all new distances
DistanceMetricClassifier now supports changing metrics when using the score() method
Added clips to make sure input for log and sqrt is appropriate.
Improved docs
Other minor formatting changes Full Changelog: v0.1.5...v0.2.0