Bootstrap.jl News and Changes
Version 2.4.0
Changes
- Extend compatibility for
Statistics
,StatsBase
,StatsModels
andRandom
. - Clarify return value of
confint
. Contributed by Hector Perez (@hdavid16) (#95). - Update doc strings. Contributed by Hector Perez (@hdavid16) (#95).
- Update documentation.
Version 2.3.3
Changes
- Extend supported versions of
Distributions
to v0.25.
Version 2.3.2
Changes
- Extend supported versions of
DataFrames
to v1.0. Contributed by Bogumił Kamiński (@bkamins) (#83). - Update DOI badge.
Version 2.3.1
Changes
- Improve documentation deployment.
Version 2.3.0
Changes
- Performance improvements through parameterized types. Contributed by David Widmann (@devmotion) (#66).
- Specialized Jack-Knife estimator. Contributed by David Widmann (@devmotion) (#67).
- Switch to Github Actions for testing and documentation generation (#75).
- Update links to external webpages in the documentation (#75).
- Extend support to
Distributions
v0.24 (#76).
Version 2.2.1
Changes
- Use upper bounds for dependencies. Contributed by Milan Bouchet-Valat (@nalimilan) (#74).
Version 2.2.0
Changes
- Support StatsModels v0.6 with
FormulaTerm
. Contributed by Dave Kleinschmidt (@kleinschmidt) (#57). - Use
RDatasets
instead of internal data sets. Contributed by David Widmann (@devmotion) (#60). - Fix deprecations in
DataFrames
andGLM
(#62). - Increase compatibility with package management in newer julia versions (#61).
Version 2.1.0
Changes
- Update continuous integration configuration (#39)
- Widen the signature for
draw!
: No longer restrict source and destinationAbstractVectors
to be of the same type, and switch to the more generalAbstractMatrix
type. Contributed by Dave Kleinschmidt (@kleinschmidt) (#43). - Improve error handling of
BCaConfInt
if all the input data has the same value. Contributed by Rory Finnegan (@rofinn) (#40). - Add Project.toml and Manifest.toml. Contributed by Colin Bowers (@colintbowers) (#48).
- Restrict compatible versions of the
StatsModels
dependency. Contributed by Mike Molignano (@mmolignano) (#53).
Version 2.0.1
Changes
- Fix
Documenter
version to v0.19 for building the documentation (#35) - Fix
convert
deprecation inDataFrames
, bump required version to v0.16.0
Version 2.0.0
Interface changes
- Put the function argument first in the signatures (#26): Changes the bootstrap signature to have the function argument come first:
bootstrap(statistic::Function, data, sampling)
. The old syntax has been deprecated and will be removed in a future version. - Integrate better with StatsBase function names (#24): Integrates with the naming of the equivalent functions in StatsBase.jl: Renames ci to confint and se to stderror. The old function names have been deprecated, and will be supported until the next major release. The motivation for these change is outlined in #23.
Changes
- Build the html documentation directly with Documenter (#28, #25)
- Support julia v0.7 and v1.0, drop support for julia v0.6 (#33)
Support
- Requires julia 0.7 or v1.0
Version 1.2.0
New features
- Maximum Entropy bootstrapping for dependent and non-stationary datasets (
MaximumEntropySampling
), contributed by Rory Finnegan (@rofinn).
Changes
- Transition tests from
FactCheck
toBase.Test
. - Introduces
StatsModels
as a new dependency which is now the new home ofFormula
s. - Use vectorized function calls for
Distribution
functions. - Removes the dependency on
Compat
.
Support
- Requires julia 0.6 or newer
- Drops support for julia 0.5
Version 1.1.0
Changes
- Adds compatibility with julia 0.6 development versions, inc- Supports formula macros for DataFrame formulas (required for using
DataFrames
v0.9.0 or newer) - Fixes convergence errors of GLM models in parametric bootstrap test cases
- Modernise automated test and coverage setup (also thanks to Tony Kelman for contributions)
- Improves the readme (also thanks to Patrick Kofod Mogensen for contributions)
luding automated tests.
- Introduces
Compat
as a new dependency - Supports formula macros for DataFrame formulas (required for using
DataFrames
v0.9.0 or newer) - Fixes convergence errors of GLM models in parametric bootstrap test cases
- Modernise automated test and coverage setup (also thanks to Tony Kelman for contributions)
- Improves the readme (also thanks to Patrick Kofod Mogensen for contributions)
Support
- Requires julia 0.5 or 0.6
- Drops support for julia 0.4
Version 1.0.0
Version 0.3.3
Type compatibility with julia 0.4
Test code in readme as part of the unit tests
Script to rebuild notebook automatically
Unified number of observations
nobs
interface with testsCalculate number of exact bootstrap samples explictly
Doc notebooks: Cleanup
Version 0.3.2
Official support for julia 0.4
Precompilation in julia 0.4
DOI for release versions, making it possible to properly cite the software
Improved
show
method forBootstrapSample
andBootstrapCI
Version 0.3.1
Sample quantiles with Gaussian interpolation
Generalize integer types
Windows integration and build service
Codecov
Version 0.3.0
Improved and more consistent documentation, using simple docstrings
Data summary in show methods, instead of displaying the entire data set
Reference list as part of the documentation
Version 0.2.0
Documentation of the package and the functions on readthedocs.org
Require julia v0.3, excluding prereleases
Version 0.1.1
Version 0.1.0
Major changes:
Simplify travis
Readme: Package status and logo
Interface: Individual functions for
boot
andci
Support
Array
andDataFrame
inputRequire:
DataFrames
andDocile
Doc: Workflows in notebooks
Doc: Function documentation with Docile
Extend balanced bootstrap functionality
BCa CI with jackknife
Parametric
Bootstrap*
typesGeneralized sampling methods
IO in
show
methodsTest: Doctest based on
Docile
Test for new functionality
Version 0.0.3
Test cases for
Distributions
dependenciesModular tests
Remove unused plotting dependencies
Version 0.0.2
- Normal CI in
boot_ci
, with test cases
Version 0.0.1
Initial release