UncertainData v0.16.0

Diff since v0.15.0

Merged pull requests: - allow current IntervalArithmetic version (#171) (@JeffreySarnoff)

Scientific Software - Peer-reviewed - Julia
Published by github-actions[bot] over 4 years ago

UncertainData v0.15.0

Diff since v0.14.0

Closed issues: - Tests sometimes fail? (#143) - Tag v0.14 (#168)

Merged pull requests: - CompatHelper: bump compat for "Distributions" to "0.25" (#170) (@github-actions[bot])

Scientific Software - Peer-reviewed - Julia
Published by github-actions[bot] over 4 years ago

UncertainData v0.14.0

Diff since v0.13.1

Closed issues: - Register v0.13.1 (#166)

Merged pull requests: - Fix stochastic bugs when sampling using StrictlyIncreasing constraints (#167) (@kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by github-actions[bot] almost 5 years ago

UncertainData v0.13.1

Diff since v0.13.0

Closed issues: - Register v0.13 (#164)

Merged pull requests: - CompatHelper: bump compat for "IntervalArithmetic" to "0.18" (#165) (@github-actions[bot])

Scientific Software - Peer-reviewed - Julia
Published by github-actions[bot] almost 5 years ago

UncertainData v0.13.0

Diff since v0.12.0

Closed issues: - Register v0.12 (#162)

Merged pull requests: - Bug fix sequential resampling (#163) (@kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by github-actions[bot] almost 5 years ago

UncertainData v0.12.0

Diff since v0.11.0

Closed issues: - Register v0.11.0 (#160)

Merged pull requests: - Bugfix in resample UncertainIndexValueDataset (#161) (@kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by github-actions[bot] almost 5 years ago

UncertainData v0.11.0

Diff since v0.10.4

Closed issues: - Register v0.10.4 (#157)

Merged pull requests: - Newer version of Plots uses xerror/yerror instead of xerr/yerr (#159) (@kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by github-actions[bot] almost 5 years ago

UncertainData v0.10.4

Diff since v0.10.3

Closed issues: - Register v0.10.3 (#141)

Merged pull requests: - CompatHelper: bump compat for "Unitful" to "0.18" (#142) (@github-actions[bot]) - CompatHelper: bump compat for "RecipesBase" to "0.8" (#147) (@github-actions[bot]) - Install TagBot as a GitHub Action (#148) (@JuliaTagBot) - CompatHelper: bump compat for "Distributions" to "0.23" (#149) (@github-actions[bot]) - CompatHelper: bump compat for "StatsBase" to "0.33" (#150) (@github-actions[bot]) - CompatHelper: bump compat for "IntervalArithmetic" to "0.17" (#151) (@github-actions[bot]) - CompatHelper: bump compat for "HypothesisTests" to "0.10" (#152) (@github-actions[bot]) - Tagbot update (#156) (@kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by github-actions[bot] about 5 years ago

UncertainData.jl - v0.10.3

v0.10.3 (2019-11-19)

Diff since v0.10.2

Closed issues:

Register v0.10.2 (#137)
Control how bins are represented when using BinnedResampling (#135)
Register v0.10.1 (#134)

Merged pull requests:

Release 0.10.3 (#138) (kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.10.2

UncertainData.jl v0.10.2

Improvements

The user can now control how each bin is represented when using BinnedResampling. One can now provide BinnedResampling{UncertainScalarKDE}, BinnedResampling{UncertainScalarPopulaton} or BinnedResampling{RawValues}.
Explicit bin methods for binning both scalar valued data and uncertain data.

Documentation

Added documentation for binning methods.
Improved documentation for UncertainScalarKDE.

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.10.1

v0.10.1 (2019-11-13)

Diff since v0.10.0

Closed issues:

Register v0.9.3 (#127)

Merged pull requests:

Release 0.10.1 (#133) (kahaaga)
Fix typo in docs (#132) (kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.10.0

UncertainData.jl v0.10.0

Improvements

The resample family of methods for vectors now dispatches on AbstractVectors, which allows more flexibility. Now, for example LArrays from LabelledArrays.jl also can be resampled.
Relax resample(x::Real) to resample(x::Number).

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - Release v0.9.3

UncertainData.jl v0.9.3

dimension is no longer exported.

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.9.2

UncertainData.jl v0.9.2

New features

Added SensitivityTests module defining the abstract type SensitivityTest.

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.9.1

v0.9.1 (2019-11-05)

Diff since v0.9.0

Closed issues:

Register v0.8.2 (#118)

Merged pull requests:

Release 0.9.1 (#122) (kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.9.0

Release v0.9.0

New features

Added interpolate_and_bin function.
Added InterpolateAndBin type.
Added resample(inds, vals, resampling::InterpolateAndBin{Linear}) method, which interpolates and bins inds and vals onto an interpolation grid, then bins and summarises the bins. Returns the binned values.
Added resample(x::AbstractUncertainIndexValueDataset, resampling::InterpolateAndBin{Linear}) method. Draws a single realisation of both the indices and values of x and orders them sequentially according to the indices (assuming independent points). Then, interpolate, bin and summarise bins.
Added bin and bin! functions.
Added bin_mean function.
Added fill_nans, fill_nans! and interpolate_nans functions for dealing with data containing NaNs.
Added findall_nan_chunks function for identifying consecutive NaNs in a dataset.
Added RandomSequences resampling scheme.

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - Release v0.8.4

Release v0.8.4

Update DOI for paper for JOSS publication

Scientific Software - Peer-reviewed - Julia
Published by kahaaga over 6 years ago

UncertainData.jl - Release v0.8.3

Release v0.8.3

A release to trigger Zenodo archiving.

Scientific Software - Peer-reviewed - Julia
Published by kahaaga over 6 years ago

UncertainData.jl - Release v0.8.2

Uncertaindata.jl v0.8.2

New features

Added resample method for BinnedWeightedResampling scheme.
Added AbstractBinnedResampling for binned resamplings.
Added AbstractBinnedUncertainValueResampling abstract type for binnings where the values in each bin is represented by an uncertain value. BinnedResampling and BinnedWeightedResampling are subtypes AbstractBinnedUncertainValueResampling.
Added AbstractBinnedSummarisedResampling abstract type for binnings where the values in each bin are summarised to a single value. BinnedMeanResampling and BinnedMeanWeightedResampling are subtypes AbstractBinnedResampling.

Improvements

Added more tests for binned resampling schemes.

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - Release v0.8.1

Uncertaindata.jl v0.8.1

New features

Added UncertainValueDataset, UncertainIndexDataset, and UncertainIndexValueDataset constructors for vectors of numbers (they get converted to CertainValues).

Bug fixes

rand(x::CertainValue, n::Int) now returns a length-n array with x repeated n times.

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - Release v0.8.0

Uncertaindata.jl v0.8.0

New functionality

Added binned resampling methods that uses BinnedResampling and BinnedMeanResampling schemes.
- resample(::AbstractUncertainIndexValueDataset, ::BinnedResampling
- resample(::AbstractUncertainIndexValueDataset, ::BinnedMeanResampling

Bug fixes

Fixed bug where resample! method for vectors and tuples of uncertain values didn't return the expected result.

Improvements

Improved resample! docs.

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.7.0

v0.7.0 (2019-10-22)

Diff since v0.6.0

Closed issues:

Mathematical operations on uncertain value datasets should return uncertain value datasets (#110)
Register v0.6.0 (#109)
Add method resample\(x::AbstractUncertainIndexValueDataset, resampling::SequentialResampling` (#107)
Add SequentialResampling type (#106)
UncertainScalarKDE and UncertainValue constructors for Vector{Array{\<:Real, 0}} (#105)
add resample! for in-place resampling (#97)
Bug when resampling fitted distributions (#90)
Mathematical operations (#28)

Merged pull requests:

Release 0.7.0 (#111) (kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.6.0

v0.6.0 (2019-10-13)

Diff since v0.5.1

Closed issues:

Register v0.5.1 (#104)
Register v0.5.0 (#102)

Merged pull requests:

Release 0.6.0 (#108) (kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.5.1

UncertainData.jl v0.5.1

Bug fixes

Strictly increasing or decreasing sequences were not always possible to construct when using CertainValues, because TruncateRange instances with equal minimum and maximum was constructed (not possible). It is now possible to resample with sequential constraints even with the StrictlyIncreasing and StrictlyDecreasing constraints.

Details

v0.5.1 (2019-10-10)

Diff since v0.5.0

Closed issues:

Tutorial explaining the difference between point-estimates for a statistic and estimates of the statistic for uncertain datasets (#92)
Allow population members to consist of distributions? (#54)

Merged pull requests:

Release v0.5.1 (#103) (kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - Release v0.5.0

UncertainData.jl v0.5.0

Breaking changes

To allow easier multiple dispatch, the indices field of a UncertainIndexValueDataset is now always an instance of a subtype of AbstractUncertainIndexDataset. The values field of a UncertainIndexValueDataset is now always an instance of a subtype of AbstractUncertainValueDataset.

New functionality

Experimental support for nested populations.
Added point-estimators for single uncertain values:
1. harmmean(x::AbstractUncertainValue, n::Int)
2. geomean(x::AbstractUncertainValue, n::Int)
3. kurtosis(x::AbstractUncertainValue, n::Int; m = mean(x))
4. moment(x::AbstractUncertainValue, k, n::Int, m = mean(x))
5. percentile(x::AbstractUncertainValue, p, n::Int)
6. renyientropy(x::AbstractUncertainValue, α, n::Int)
7. rle(x::AbstractUncertainValue, n::Int)
8. sem(x::AbstractUncertainValue, n::Int)
9. skewness(x::AbstractUncertainValue, n::Int; m = mean(x))
10. span(x::AbstractUncertainValue, n::Int)
11. summarystats(x::AbstractUncertainValue, n::Int)
12. totalvar(x::AbstractUncertainValue, n::Int)
Added statistical estimators for pairs of uncertain values:
1. cov(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int; corrected::Bool = true)
2. cor(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
3. countne(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
4. counteq(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
5. corkendall(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
6. corspearman(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
7. maxad(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
8. meanad(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
9. msd(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
10. psnr(x::AbstractUncertainValue, y::AbstractUncertainValue, maxv, n::Int)
11. rmsd(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int; normalize = false)
12. sqL2dist(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
13. crosscor(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int; demean = true)
14. crosscov(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int; demean = true)
15. gkldiv(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
16. kldivergence(x::AbstractUncertainValue, y::AbstractUncertainValue, n::Int)
Added UncertainValue constructor for distribution instances.
Added UncertainValue constructor for (potentially nested) truncated distribution instances.
Implemented resample methods for NTuples of uncertain values.
Added resample(f::Function, n::Int, x::AbstractUncertainValue, args...; kwargs...)method for easy evaluation of point-estimates for single uncertain values.
Added support for Measurement instances from Measurements.jl. These are treated as uncertain values represented by normal distibutions. Hence, they are given no extra treatment and error propagation is done by resampling, not by exact methods.
The uncertain value type UncertainScalarPopulation may now not only have real-valued scalars as elements of the population. It can now have uncertain values as members of the population!
Resampling implemented for UncertainScalarPopulation so that we can also sample population members that are uncertain values.
Implemented iteration for UncertainScalarPopulation.

Improvements

Improved subtyping for theoretical distributions.
Removed redundant resample methods for the UncertainDataset type. UncertainDataset is a subtype of AbstractUncertainValueDataset and has no special behaviour beyond that implemented for the abstract type, so now we just rely on multiple dispatch here.

Documentation

Improved documentation statistical methods.
Other minor documentation improvements.
Improved documentation for TruncateStd.

Bug fixes

Fixed error in show method for AbstractUncertainValue. Not subtypes of AbstractUncertainValue has the distributions field, so that is now removed from the show method.

Details

v0.5.0 (2019-10-09)

Diff since v0.4.0

Closed issues:

promotion when resampling (#99)
Add single-estimate methods to cor and cov for uncertain datasets (#95)
Improve docs for cor and cov for uncertain datasets (#94)
cor and cov point-estimates (#93)
cor and cov for UncertainIndexValueDataset (#91)
resample n-tuples (#89)
Promote scalar values to CertainValue (#88)
Register v0.4.0 (#84)
Register v0.3.0 (#81)
identity resample for real-valued vectors (#79)
Resample Vector{AbstractUncertainValue} using SamplingConstraints (#76)

Merged pull requests:

Release 0.5.0 (#86) (kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.4.0

UncertainData.jl v0.4.0

New functionality

Introduce an abstract resampling type AbstractUncertainDataResampling for this package pending the implementation of AbstractResampling in StatsBase.jl.
Added ConstrainedResampling resampling scheme.
Resample vectors of uncertain values without constraints. Syntax:
1. resample(::Vector{<:AbstractUncertainValue} for single draws.
2. resample(::Vector{<:AbstractUncertainValue}, ::Int} for multiple draws.
Resample vectors of uncertain values with constraint(s) multiple times. Syntax:
1. resample(::Vector{<:AbstractUncertainValue}, ::Union{SamplingConstraint, Vector{<:SamplingConstraint}} for single draws.
2. resample(::Vector{<:AbstractUncertainValue}, ::Union{SamplingConstraint, Vector{<:SamplingConstraint}}, ::Int for multiple draws.

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.3.0

v0.3.0 (2019-09-12)

Diff since v0.2.3

Closed issues:

Register v0.2.3 (#73)

Merged pull requests:

Update version (#82) (kahaaga)
Release v0.3.0 (#80) (kahaaga)
Removed manifest from repo and add to .gitignore (#78) (kahaaga)
Update README.md (#77) (kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.2.3

UncertainData.jl v0.2.3

Improvements

Added input validation when initialising TruncateQuantiles, TruncateRange and TruncateStd.
Separate parameters types for TruncateQuantiles and TruncateRange, so one can do for example TruncateRange(1, 8.0), instead of having to promote to Float64.
Added validation for distribution truncation when resampling.

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.2.2

UncertainData.jl v0.2.2

New functionality and syntax changes

Resampling vectors consisting of uncertain values (done in #61)

resample(uvals::Vector{AbstractUncertainValue}, n::Int) is now interpreted as "treat uvals as a dataset and sample it n times". Thus, it now behaves as resample(AbstractUncertainDataset, n::Int), returning n vectors of length length(uvals), where the i-th element is a unique draw of uvals[i].
resample_elwise(uvals::Vector{AbstractUncertainValue}, n::Int) takes over the role as "sample uvals element-wise and n times for each element". Returns a vector of length length(uvals), where the i-th element is a n-element vector of unique draws of uvals[i].

Resampling with subtypes of `AbstractUncertainValueDataset`

Currently, this affects the generic UncertainDatasets, as well as the specialized UncertainIndexDatasets and UncertainValueDatasets.

resample_elwise(uvd::AbstractUncertainValueDataset, n::Int) is now interpreted as "draw n realisations of each value in uvd". Returns a vector of length length(uvals) where the i-th element is a n-element vector of unique draws of uvals[i]. This works for UncertainDatasets, UncertainIndexDatasets, and UncertainValueDatasets.
resample_elwise(uvd::AbstractUncertainValueDataset, constraint::Union{SamplingConstraint, Vector{SamplingConstraint}}, n::Int) is now interpreted as "draw n realisations of each value in uvd, subjecting each value in uvd to some sampling constraint(s) during resampling". Returns a vector of length length(uvals) where the i-th element is a n-element vector of unique draws of uvals[i], where the support of uvals[i] has been truncated by the provided constraint(s).

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - Release v0.2.1

UncertainData.jl v0.2.1

New functionality

merge(uvals::Vector{<:AbstractUncertainValue}; n = 1000) now makes it possible to combine many uncertain values of different into one uncertain value represented by a kernel density estimate. This is achieved by resampling each of the values n times, then pooling the draws and estimating a total distribution using KDE.
merge(uvals::Vector{<:AbstractUncertainValue}; weights::Weights n = 1000), merge(uvals::Vector{<:AbstractUncertainValue}; weights::AnalyticalWeights n = 1000) and merge(uvals::Vector{<:AbstractUncertainValue}; weights::ProbabilityWeights n = 1000) merges uncertain values by resampling them proportionally to weights, then pooling the draws and performing KDE. These are all functionally equivalent, but implementations for different weights are provided for compatibility with StatsBase.
merge(uvals::Vector{<:AbstractUncertainValue}; weights::FrequencyWeights n = 1000) merges uncertain values by sampling them according to the number of samples provided with weights.

Bug fixes

resample didn't work for UncertainIndexDatasets due to the data being stored in the indices field, not the values field as for other subtypes of AbstractUncertainValueDataset. This is now fixed.

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.2.0

v0.2.0 (2019-08-13)

Diff since v0.2.1

Merged pull requests:

Release.0.2.0 (#51) (kahaaga)
Release 0.2.0 (#50) (kahaaga)
Release-0.2.1 (#49) (kahaaga)

Scientific Software - Peer-reviewed - Julia
Published by julia-tagbot[bot] over 6 years ago

UncertainData.jl - v0.1.8

UncertainData.jl v0.1.8

Bug fixes

Added missing package dependencies which were not caught by CI. These caused installation of the package using Julia's package manager to fail.

Scientific Software - Peer-reviewed - Julia
Published by kahaaga about 7 years ago

UncertainData.jl - v0.1.7

UncertainData.jl v0.1.7

New functionality

UncertainIndexValueDatasets can now be constructed from vectors of uncertain values. To do so, provide a vector of uncertain values for the indices, and the same for the values, e.g. UncertainIndexValueDataset([idx1, idx2], [val1, val2]).
Index-value dataset realizations can now be interpolated on a regular grid. To do so, provide resample with a RegularGrid instance.

Improvements

minima and maxima now returns the global minimum for a dataset instead of a vector of elementwise minima and maxima.
Removed/merged redundant method definitions and multiple imports of the same files causing definitions to be overwritten and multiple warnings statements to be printed when loading the package.

Bug fixes

Fixed non-critical indexing bug for uncertain index-value datasets.
Changed package UUID because the old one did not work on CI when tagging the package.

Scientific Software - Peer-reviewed - Julia
Published by kahaaga about 7 years ago

UncertainData.jl - v0.1.6

UncertainData.jl v0.1.6

New functionality

Implemented sequential sampling constraints StrictlyIncreasing and StrictlyDecreasing for UncertainIndexValueDatasets.
Added UncertainScalarPopulation type, representing vectors of values that should be sampled according to a vector of probabilities.

Improvements

Improved documentation for CertainValues.
Added documentation for UncertainScalarPopulation.
Added UncertainScalarPopulation to uncertain value overview list in the documentation.
Fixed duplicate docs for cot, cotd, coth and added missing acot, acotd, acoth docs.
Shortened and updated main documentation page with more links.

Bug fixes

Import Base functions properly when defining CertainValue, so that no unexpected behaviour is introduced.
Fixed links in documentation that pointed to the wrong locations.
Remove model resampling docs which was not supposed to be published until the functionality is properly implemented.

Scientific Software - Peer-reviewed - Julia
Published by kahaaga about 7 years ago

UncertainData.jl - v0.1.5

UncertainData.jl v0.1.5

New functionality

Added CertainValue <: AbstractUncertainValue type to represent scalars without any uncertainty. Even though a scalar is not uncertain, we'll define it as subtype of AbstractUncertainValue to treat certain values alongside uncertain values in datasets.
Added plot recipe for CertainValues. They are just plotted as regular points.
Added method resample(Vector{AbstractUncertainValue}) for resampling vectors of uncertain values. Operates element-wise, just as for an uncertain dataset.
Added an abstract type SequentialSamplingConstraint to separate sequential constraints from general constraints that might be applied before resampling according to the sequential constraints.
Added abstract type (OrderedSamplingAlgorithm) and composite types (StartToEnd, EndToStart, MidpointOutwards, ChunksForwards, ChunksBackwards) which indicates how to sample sequential realizations when resampling an uncertain dataset. Only StartToEnd is used at the moment.
Added abstract type SequentialSamplingConstraint which is the supertype for all sequential constraints.
Added function to check if strictly increasing sequences through an uncertain dataset exist: strictly_increasing_sequence_exists(udata::AbstractUncertainValueDataset.
Added function to check if strictly decreasing sequences through an uncertain dataset exist: strictly_increasing_sequence_exists(udata::AbstractUncertainValueDataset.
Added the StrictlyIncreasing{T} where {T<:OrderedSamplingAlgorithm} sequential constraint for resampling uncertain datasets.
Added the StrictlyDecreasing{T} where {T<:OrderedSamplingAlgorithm} sequential constraint for resampling uncertain datasets.
Added resampling methods

1. `resample(udata, sequential_constraint::StrictlyIncreasing{T} where {T <: StartToEnd}`
2. `resample(udata, sequential_constraint::StrictlyDecreasing{T} where {T <: StartToEnd}`
3. `resample(udata, constraint::SamplingConstraint, sequential_constraint::StrictlyIncreasing{T} where {T <: StartToEnd}`
4. `resample(udata, constraint::SamplingConstraint, sequential_constraint::StrictlyDecreasing{T} where {T <: StartToEnd}`
5. `resample(udata, constraint::Vector{SamplingConstraint}, sequential_constraint::StrictlyIncreasing{T} where {T <: StartToEnd}`
6. `resample(udata, constraint::Vector{SamplingConstraint}, sequential_constraint::StrictlyDecreasing{T} where {T <: StartToEnd}`

Improvements

Added documentation on sequential constraints, clearly separating it from the general constraints.

Scientific Software - Peer-reviewed - Julia
Published by kahaaga about 7 years ago

UncertainData.jl - v0.1.4

UncertainData.jl v0.1.4

Breaking changes

Elementary operations for (scalar, uncertain_value), (uncertain_value, scalar) and (uncertain_value, uncertain_value) pairs now returns an uncertain value instead of a vector of resampled realizations. The default behaviour is to perform a kernel density estimate over the vector of results of the element-wise operations (which was previously returned without representing it as an uncertain value).

New functionality

Implemented constraints for datasets that have already been constrained. constrain(udata::ConstrainedDataset, s::SamplingConstraint) will now return another ConstrainedDataset. The same applies for ConstrainedIndexDataset and ConstrainedValueDataset.
Added maximum(Vector{AbstractUncertainValue}) and minimum(Vector{AbstractUncertainValue}) methods.
Added plot recipe for Vector{AbstractUncertainValue}s. Behaves just as plotting an uncertain dataset, assuming an implicit indices 1:length(v). Error bars may be tuned by providing a second argument of quantiles to plot, e.g. plot(v, [0.2, 0.8] gives error bars covering the 20th to 80th percentile range of the data.

Improvements

Added documentation for StrictlyIncreasing and StrictlyDecreasing sampling constraints.
Added show function for AbstractUncertainIndexDataset. show errored previously, because it assumed the default behaviour of AbstractUncertainValueDataset, which does not have the indices field.

Bug fixes

Fixed bug when resampling an uncertain dataset using the NoConstraint constraint, which did not work to due to a reference to a non-existing variable.
Fixed test bug where when resampling an uncertain value with the TruncateStd sampling constraint, the test compared the result to a fixed scalar, not the standar deviation of the value. This sometimes made the travis build fail.

Scientific Software - Peer-reviewed - Julia
Published by kahaaga about 7 years ago

UncertainData.jl - v0.1.3

UncertainData.jl v0.1.3

Allow both the indices and values fields of UncertainIndexValueDataset to be any subtype of AbstractUncertainValueDataset. This way, you don't have to use an index dataset type for the indices if not necessary.
Improved documentation for UncertainIndexDataset, UncertainValueDataset, UncertainDataset and UncertainIndexValueDataset types and added an overview page in the documentation to explain the difference between these types.
Added an overview section for the resampling documentation.
Cleaned and improved documentation for uncertain values.
Added separate documentation for the uncertain index dataset type.
Added separate documentation for the uncertain value dataset type.
Improved documentation for the generic uncertain dataset type
Merged documentation for sampling constraints and resampling.
Added missing documentation for the sinc, sincos, sinpi, cosc and cospi trig functions.

Scientific Software - Peer-reviewed - Julia
Published by kahaaga about 7 years ago

UncertainData.jl - v0.1.2

UncertainData.jl v0.1.2

Support elementary mathematical operations (+, -, * and /) between arbitrary uncertain values of different types. Also works with the combination of scalars and uncertain values. Because elementary operations should work on arbitrary uncertain values, a resampling approach is used to perform the mathematical operations. This means that all mathematical operations return a vector containing the results of repeated element-wise operations (where each element is a resampled draw from the furnishing distribution(s) of the uncertain value(s)). The default number of realizations is set to 10000. This allows calling uval1 + uval2 for two uncertain values uval1 and uval2. If you need to tune the number of resample draws to n, you need to use the +(uval1, uval2, n) syntax (similar for the operators). In the future, elementary operations might be improved for certain combinations of uncertain values where exact expressions for error propagation are now, for example using the machinery in Measurements.jl for normally distributed values.
Support for trigonometric functions added (sin, sind, sinh, cos, cosd, cosh, tan, tand, tanh, csc, cscd, csch, csc, cscd, csch, sec, secd, sech, cot, cotd, coth, sincos, sinc, sinpi, cosc, cospi). Inverses are also defined (asin, asind, asinh, acos, acosd, acosh, atan, atand, atanh, acsc, acscd, acsch, acsc, acscd, acsch, asec, asecd, asech, acot, acotd, acoth). Beware: if the support of the funishing distribution for an uncertain value lies partly outside the domain of the function, you risk encountering errors. These also use a resampling approach, using 10000 realizations by default. Use either the sin(uval) syntax for the default, and sin(uval, n::Int) to tune the number of samples.
Support non-integer multiples of the standard deviation in the TruncateStd sampling constraint.
Fixed bug in resampling of index-value datasets, where the n arguments wasn't used.
Bugfix: due to StatsBase.std not being defined for FittedDistribution instances, uncertain values represented by UncertainScalarTheoreticalFit instances were not compatible with the TruncateStd sampling constraint. Now fixed!
Improved documentation for resampling for UncertainIndexValueDatasets. Now shows the documentation for the main methods, as well as examples of how to use different sampling constraints for each individual index and data value.
Improved documentation for resampling for UncertainDatasets. Now shows the documentation for the main methods.
Added missing resample(uv::AbstractUncertainValue, constraint::TruncateRange, n::Int) method.

Scientific Software - Peer-reviewed - Julia
Published by kahaaga about 7 years ago

UncertainData.jl - v0.1.1

UncertainData.jl v0.1.1

Indexing implemented for UncertainIndexValueDataset.
Resampling implemented for UncertainIndexValueDataset.
Uncertain values and uncertain datasets now support minimum and maximum.
support(uv::AbstractUncertainValue) now always returns an interval from IntervalArithmetic.jl
support_overlap now computes overlaps also for fitted theoretical distributions.
Added more plotting recipes.
All implemented uncertain data types now support resampling.
Improved general documentation. Added a reference to Measurements.jl and an explanation for the differences between the packages.
Improved resampling documentation with detailed explanation and plots.

Scientific Software - Peer-reviewed - Julia
Published by kahaaga about 7 years ago

UncertainData.jl - Initial release

Scientific Software - Peer-reviewed - Julia
Published by kahaaga about 7 years ago

Recent Releases of UncertainData.jl

UncertainData.jl - v0.16.0

UncertainData v0.16.0

UncertainData.jl - v0.15.0

UncertainData v0.15.0

UncertainData.jl - v0.14.0

UncertainData v0.14.0

UncertainData.jl - v0.13.1

UncertainData v0.13.1

UncertainData.jl - v0.13.0

UncertainData v0.13.0

UncertainData.jl - v0.12.0

UncertainData v0.12.0

UncertainData.jl - v0.11.0

UncertainData v0.11.0

UncertainData.jl - v0.10.4

UncertainData v0.10.4

UncertainData.jl - v0.10.3

v0.10.3 (2019-11-19)

UncertainData.jl - v0.10.2

UncertainData.jl v0.10.2

Improvements

Documentation

UncertainData.jl - v0.10.1

v0.10.1 (2019-11-13)

UncertainData.jl - v0.10.0

UncertainData.jl v0.10.0

Improvements

UncertainData.jl - Release v0.9.3

UncertainData.jl v0.9.3

UncertainData.jl - v0.9.2

UncertainData.jl v0.9.2

New features

UncertainData.jl - v0.9.1

v0.9.1 (2019-11-05)

UncertainData.jl - v0.9.0

Release v0.9.0

New features

UncertainData.jl - Release v0.8.4

Release v0.8.4

UncertainData.jl - Release v0.8.3

Release v0.8.3

UncertainData.jl - Release v0.8.2

Uncertaindata.jl v0.8.2

New features

Improvements

UncertainData.jl - Release v0.8.1

Uncertaindata.jl v0.8.1

New features

Bug fixes

UncertainData.jl - Release v0.8.0

Uncertaindata.jl v0.8.0

New functionality

Bug fixes

Improvements

UncertainData.jl - v0.7.0

v0.7.0 (2019-10-22)

UncertainData.jl - v0.6.0

v0.6.0 (2019-10-13)

UncertainData.jl - v0.5.1

UncertainData.jl v0.5.1

Bug fixes

Details

v0.5.1 (2019-10-10)

UncertainData.jl - Release v0.5.0

UncertainData.jl v0.5.0

Breaking changes

New functionality

Improvements

Documentation

Bug fixes

Details

v0.5.0 (2019-10-09)

UncertainData.jl - v0.4.0

UncertainData.jl v0.4.0

New functionality

UncertainData.jl - v0.3.0

v0.3.0 (2019-09-12)

UncertainData.jl - v0.2.3

UncertainData.jl v0.2.3

Resampling with subtypes of `AbstractUncertainValueDataset`