Skip to content

Latest commit

 

History

History
170 lines (98 loc) · 4.57 KB

CHANGELOG.md

File metadata and controls

170 lines (98 loc) · 4.57 KB
Raw

v0.11.5 (Upcoming Release)

  • Initial implementation of the robust hat matrix regression estimator
  • Add more test to robust hat matrix regression estimator
  • Introduce view in LTS and LMS.

v0.11.4

  • More explicit return types, drop Dict with Dict{String, Any} or Dict{String, Vector}
  • Add Julia v.1.10 to GitHub actions
  • Initial attempt to reduce memory allocations in lts(), lms(), hadi92(), hadi94(), hs93(), robcov()
  • Replace @assert macro with throw(ErrorException()) in whole code
  • depestregression returns Dict instead of a vector of betas like other regression methods.
  • summary() throws ErrorException rather than simply prompting with @error macro.
  • robcov doesn't use try and catch any more.
  • Replace sortperm with sortperm! in mve.
  • Set number of iterations to minimum([5 * p, 3000]) of LTS as in the R implementation. This reduces the time required 3x.

v0.11.3

  • Minor fixes

v0.11.2

  • Optional eps and maxiter parameters in iterateCSteps() in LTS.
  • Replace ols(X[indices, :], y[indices]) with X[...] \ y[...] in highly computational ones such like LTS, LMS, and LTA.

v0.11.1

  • Concrete types of X and y changed to AbstractMatrix{Float64} and AbstractVector{Float64}
  • Change function signatures from ::Function to ::F where {F <: Function}
  • Increase test coverage

v0.11.0

  • Deepest Regression Estimator added.

v0.10.2

  • mahalanobisSquaredBetweenPairs() return Union{Nothing, Matrix} depending on the determinant of the covariance matrix
  • mahalanobisSquaredMatrix() returns Union{Nothing, Matrix} depending on the determinant of the covariance matrix
  • import in DataImages fixed.
  • Array{Float64, 1} is replaced by Vector{Float64}.
  • Array{Float64, 2} is replaced by Matrix{Float64}.
  • Use of try/catch reduced, many tries were depending on singularities.

v0.10.1

  • Update compatibility with Clustering and StatsModels

v0.10.0

  • LMS returns betas rather than coefs
  • PrecompileTools integration for faster loading of package
  • Replace RGB{} to RGBX{} in plots

v0.9.5

  • Adopt to SnoopPrecompile for better first-time-use experiment

# v0.9.4

  • Convergency check for hs93()
  • Add earlystop option for inexact version of lta().
  • Remove cga() and cga based experimental algorithm.
  • Fix deadlock in bacon.

v0.9.3

  • ga() is now faster.
  • Implement dfbetas().
  • Separate implementations of dffit() and dffits()
  • Implement diagnose()
  • Replace GLPK with HiGHS. When n > 10000, HiGHS based lad() is now approximately 6x faster.
  • asm2000(), imon2005(), ks89(), bacon(), and smr98() now return regression coefficients calculated using the clean set of observations.
  • lts() has a new optional parameter called earlystop and it is true by default. If the objective function does not change in a predefined number of iterations, the search is stopped.
  • py95() returns vector of estimated regression coefficients using the clean observations.

v0.9.2

  • Increase test coverage
  • Add ols(setting) and wls(setting, weights = w) type method call where setting is a regression setting
  • Implement cooksoutliers() method for determining potential regression outliers using a cutoff value.
  • Update documentation
  • Implement Theil-Sen estimator for multiple regression

v0.9.1

  • Fix bchplot dependencies
  • Update README with new instructions
  • Satman2015 now returns more verbose output

v0.9.0

  • Add exact argument for LAD. If exact is true then the linear programming based exact solution is found. Otherwise, a GA based search is performed to yield approximate solutions.
  • Remove dependency of Plots.jl. If Plots.jl is installed and loaded manually, the functionality that uses Plot is autmatically loaded by Requires.jl. Affected functions are dataimage, mveltsplot, and bchplot.

v0.8.19

  • Update Satman(2013) algorithm

v0.8.18

  • Add docs for Satman's (modified) GA based LTS estimation (2012)

v0.8.17

  • Remove dependency of StatsBase

v0.8.16

  • Quantile Regression implemented

v0.8.15

  • Modularize dataimage()
  • Grouped tests
  • x === nothing style decisions replaces by isnothing(x)

v0.8.14

  • Update documentation
  • Modularization

v0.8.13

  • Removed some unused variables
  • Refactor code
  • Update docs system

v0.8.12

  • Dependency entries updated in Project.toml

v0.8.11

  • asyncmap replaced with map in lta
  • JuMP version increased

v0.8.10

  • Julia compatibality level is now 1.7
  • Update JuMP and GLPK

v0.8.9

  • LAD (Least Absolute Deviations) is now exact and uses a linear programming based model
  • Dependencies for JuMP and GLPK are added
  • Dependency for Optim removed