Release v0.3.2 

Release v0.3.1 

Release v0.3 

Release version 0.2.0 

Release version 0.1.4 

release for the submission of the paper 

refactor to scikit-matter 

Release version 0.1.2 

Release version 0.1.1 

Feat/kpcovr fitted regressor (#113) * Add shape checking utilities for coefficients of precomputed kernel regressors * Modify instantiation and fit call of KPCovR to accept pre-fitted regressors as in PCovR * Update KPCovR tests to be compatible with new regressor usage * Update PCovR example notebook to be compatible with new regressor usage * Reorganize regressor usage to pull kernel parameters directly from the regressor; use None as the default argument for the regressor * Pull alpha from the KPCovR regressor * Make regressor default argument None, assign default within __init__ * Change inversions to use least squares with singular value cutoff based on tol instead of the regularization * Compute Yhat directly from the dual coefficients * Move regressor checking to occur immediately * Add more details about pre-fitted regressors to PCovR and KPCovR documentation * Use KPCovR tolerance in matrix inversion instead of regularization * Add tests for KPCovR to cover the pre-fitted regressors * Add PCovR test to check for regressor modifications * Move default regressor assignment to fit and accept regressor params * Reorganize KPCovR regressor infrastructure * Make PCovR example compatible with new KPCovR regressor infrastructure * Add PCovR test for None regressor * Modify KPCovR tests for compatibility with new regressor infrastructure * Add KPCovR test for None regressor * Fix KPCovR docstring example * Consolidate regressor checking * Simplify tests for pre-fitted regressors * Negate KPCovR score according to sklearn guidelines