18 NEWS

001 04/05/16

  • started the whole thing

002 04/12/16

  • inserted display code
  • finished GALO example in MCA chapter
  • figured out multiple ordinal
  • added empty addals function and chapter
  • started caching computations
  • added empty introductory chapter

003 04/12/16

  • added preface and introduction (partial)
  • added cover photo

004 04/13/16

  • added text to preface
  • gifiEngine() now returns QR decompositions of the bases
  • added acknowledgements to preface

005 04/14/16

  • small corrections
  • refactoring main code
  • putting in missing data (4 options)
  • putting in monotone regression (3 options for ties)

006 04/16/16

  • completely rewritten R engine (does not work yet)
  • new code in last chapter (does not work yet)
  • rational treatment of regression and missing data
  • currently cone/subspace regression uses Dykstra algorithm
  • a Gifi is a list of gifiSets, a gifiSet is a list of gifiVariables
  • a gifiVariable stores the things that stay the same during computation
  • an xGifi is a list of xGifiSets, an xGifiSet is a list of xGifiVariables
  • an xGifiVariable stores the things that change during computation

007 04/17/16

  • more code refactoring
  • new code (in appendices) still does not work properly
  • old code still used in body of paper for computations

008 04/18/16

  • corrected some serious bugs, new gifiEngine() code now runs
  • missing data handling still needs work

009 04/18/16

  • fixed the homals() wrapper
  • added corList() utility
  • added starPlotter() utility
  • eval = FALSE for wrapper examples that do not work yet
  • added section on structures
  • cached some computations
  • readied the MCA section with new software

010 04/19/16

  • fixed the corals wrapper
  • added utilities preCorals() and preCoranals
  • many small changes
  • added CA text
  • added empty section on CCA

011 04/21/16

  • small edits
  • added some preface/introduction
  • corals example
  • missing data section (as yet empty)

012 04/22/16

  • something on missing data
  • added to todo list
  • more in preface/introduction

013 04/22/16

  • added active slot to variables
  • preparing to add aspect engine
  • start merging old and new loss functions

014 04/24/16

  • utilities for Burt table, profile frequencies, and interactive coding
  • include decode and encode in C/R
  • gifiEngine takes an object of class gifi as argument
  • homals wrapper builds an object of class gifi
  • added data weights parameter (does not do anything yet)

015 04/27/16

  • changes to correctly handle missing data
  • small edits
  • simplified notation

016 04/30/16

  • additions to intro and history

017 05/02/16

  • active and passive variables handling

018 05/03/16

  • split gifi.R from gifiEngine.R
  • gifiEngine.R just takes a gifi argument, no longer data argument
  • split off gifiTransform() from gifiEngine()
  • split off dykstra() from coneRegression()
  • more on copies and interactive coding
  • included updated code in index.Rnw

019 05/04/16

  • fixed morals(), but needs some testing
  • modified isotone() so that only non-missing data are constrained to be monotone
  • adapted coneRegression() to isotone() with missing data
  • fixed active/passive variables
  • included code for aspectEngine() – not used yet

020 05/05/16

  • implemented missing data “average” using dykstra
  • repaired princals wrapper
  • repaired NLPCA chapter

021 05/05/16

  • repaired criminals() wrapper
  • repaired CDA chapter

022 05/06/16

  • added makeColumnProduct() utility
  • small example in missing data section

023 05/07/16

  • expanded correspondence analysis chapter
  • added Glass example

024 05/08/16

  • added regressionPlotter() utility
  • added regression plots to OCA chapter

025 05/09/16

  • small improvement in makeIndicator()
  • added burt table to corals output
  • improved and explained plots in Glass example
  • incorporated preCorals() in corals()
  • added postCorals() utility
  • filled in Galton example
  • modified corals() to accept either tables or bivariables
  • modified gifiEngine() to do at least two iterations

026 05/10/16

  • put in latest code
  • worked on preface and introduction

027 05/12/16

  • added loss function history

028 05/18/16

  • started change to linear systems loss

029 05/19/16

  • introduced orthoblocks
  • described MVAOS loss

030 05/20/16

  • MVAOS system examples
  • Gifi loss as special case
  • more on orthoblocks

031 05/21/16

  • More on Gifi loss

032 05/25/16

  • makePattern() function
  • majorization algorithm

033 05/30/16

  • rewriting core code – will take a while
  • going to aspects completely
  • writing out blocks and Gifi
  • engine transfors variables and calls pattern subroutine
  • pattern subroutines make aspect gradients
  • changed title

034 06/04/16

  • edits, working the core code