Releases: bashtage/arch
Release 7.2
Release 7.1
This is a minor release that fixes one unlikely to encounter bug and improves the documentation.
Release 7.0.0
- Full compatability with NumPy 2
- Improved compatability with future changes in pandas 3.
- Increases in related minimum requirements.
Note
In order to use NumPy 2, the environment must consist of packages that
have been built against NumPy 2.0.0rc1 or later.
Release 6.3
This is a performance and compatibility release.
- Improve performance of long-run covariance estimators when numba is installed
- Add support for NumPy 2
- Improve compatibility with pandas
Release 6.2.0
This is a maintenance and bug-fix release:
- Fixed a bug that affected forecasting from
FIGARCH
models #606 - Added a performance warning when testing for unit roots in large series using a lag-length search with no-max-lag specified.
- Changed the default value of reindex to False so that forecasts will not match the input by default. Set reindex to True if this is required.
- Made from
__future__ import reindex
a no-op. - Updated notebooks to reflect best practices
Release 6.1.0
This release fixes a bug that affects forecasting when:
- using an AR model or other model with lagged values of the dependent variable; and
- data is rescaled.
Release 5.6.0
This release fixes a non-trivial bug. It is likely the final release from the 5.x branch, and has been released so that users on Python 3.8 will not be affected.
Release 6.0.1
This release only rolls back Cython from the 3.0.0 branch to the latest release 0.29.34. This change to allow arch to be built on conda-forge which does not have support for Cython 3.0.
- Rollback Cython support from 3.0.0b2+ to 0.29.34+
Release 6.0.0
This major release makes the following changes that affect the supported configurations and build system.
- Minimum supported Python is 3.9
- Bumped minimum NumPy, SciPy, pandas, statsmodels and Cython
- Removed dependence on property-cached
- Added compatibility with Cython 3
Release 5.5.0
This is a compatibility release the improves compatibility with NumPy 1.25 (unreleased) and adds initial support for pandas copy-on-write behavior. There are no new features.