Skip to content

Latest commit

 

History

History
68 lines (44 loc) · 1.86 KB

releases.md

File metadata and controls

68 lines (44 loc) · 1.86 KB

2019.2

Numpy requirements updated to be >= 1.14

New features

Bug fixes

Enhancements

  • New unit-test for rolling window nanmedian
  • Removed support for numpy < 1.14, and fixed formatting from earlier numpy versions

2019.1

Numba requirements changed to be >= 0.41 due to Windows segfaults

New features

Bug fixes

  • Fixed the numba lowering error from the scaling check for ddof not in (0, 1)

Enhancements

  • Now testing all PRs on python 3.5, 3.6 and 3.7 on Linux and Windows, and 3.7-dev on linux
  • Travis CI using Xenial instead of Trusty to get python 3.7 support

2018.1

New features

  • Exponentially Weighted Moving Average

Bug fixes

Enhancements

2017.1

New features

  • newton_raphson : Root finding using the Newton-Raphson iteration method
  • erf : error function using Abramowitz and Stegun method (maximum error: 1.5e-7)
  • correlation : Spearman and Pearson correlation coefficient functions
  • pca : Principal Component Analysis, returning the transformed data, not eigenvalues or eigenvectors.
  • Added Windows CI builds: #23
  • LU Decomposition
  • binary_search : Root finding using the bisection method
  • Documentation added using sphinx + numpydoc.
  • OLS: f_statistic
  • QR: classical and modified Gram Schmidt methods
  • Matrix inverse using adjoint method
  • Matrix determinant
  • Matrix minor (sub-matrix with one row and one column eliminated)
  • Scaling functions (standard, min_max, rank, demean, shrink off diagonals)
  • Lasso regression for orthonormal covariates (features)
  • drop_missing : helper function analogous to statsmodels missing='drop' mechanism which allows the user to evict features and observations where one or more data points is not finite such that OLS may then be performed on dense / complete data.
  • EWMA functions for 1D and 2D data sets (dense data / no NaNs)

Bug fixes

Enhancements