[Home]
[CV]
[Publications]
[Teaching]
[Software]




 


Florian Wickelmaier

Software

1 Multi-attribute probabilistic choice models
2 Multinomial processing tree (MPT) models
3 Probabilistic knowledge structures

1 Multi-attribute probabilistic choice models

R and Matlab functions for fitting and testing multi-attribute probabilistic choice models, especially the Bradley-Terry-Luce (BTL) model (Bradley & Terry, 1952; Luce, 1959), elimination-by-aspects (EBA) models (Tversky, 1972), and preference tree (Pretree) models (Tversky & Sattath, 1979).

eba package


  library(eba)
  
  data(drugrisk)
  
  eba(
    drugrisk[,,group = "male30"],
    list(c(1),
         c(2, 7),
         c(3, 7),
         c(4, 7, 8),
         c(5, 7, 8),
         c(6, 7, 8))
  )

Requires: R 2.0.0 or higher
Documentation: R package manual [pdf]
Source code: eba_1.7-0.tar.gz
Windows binary: eba_1.7-0.zip

Also available from the Comprehensive R Archive Network (CRAN).

OptiPt

Requires: Matlab 6.0 or higher
Source code: OptiPt.m

Also available from the Matlab File Exchange.

fOptiPt

Requires: Matlab 6.0 or higher + statistics toolbox
Source code: fOptiPt.m

Reference

Wickelmaier, F., & Schmid, C. (2004). A Matlab function to estimate choice model parameters from paired-comparison data. Behavior Research Methods, Instruments, and Computers, 36(1), 29-40. [pdf]

2 Multinomial processing tree (MPT) models

Fitting and testing multinomial processing tree models, a class of statistical models for categorical data. The parameters are the link probabilities of a tree-like graph and represent the latent cognitive processing steps executed to arrive at observable response categories (Batchelder & Riefer, 1999; Erdfelder et al., 2009; Riefer & Batchelder, 1988).

mpt package

  library(mpt)
  
  data(retroact)
  
  mpt(freq ~ list(
    c*r,
    (1 - c)*u^2,
    2*(1 - c)*u*(1 - u),
    c*(1 - r) +
      (1 - c)*(1 - u)^2,
    u,
    1 - u
  ),
  data = retroact[
    retroact$lists == 0,])

Requires: R 2.15.0 or higher
Documentation: R package manual [pdf]
Source code: mpt_0.4-1.tar.gz
Windows binary: mpt_0.4-1.zip

Also available from the Comprehensive R Archive Network (CRAN).

3 Probabilistic knowledge structures

Fitting and testing probabilistic knowledge structures, especially the basic local independence model (BLIM, Doignon & Flamagne, 1999), using the minimum discrepancy maximum likelihood (MDML) method.

pks package

  library(pks)
  
  data(DoignonFalmagne7)
  
  blim(
    DoignonFalmagne7$K,
    DoignonFalmagne7$N.R,
    method = "MDML"
  )

Requires: R 2.14.0 or higher
Documentation: R package manual [pdf]
Source code: pks_0.2-0.tar.gz
Windows binary: pks_0.2-0.zip

Also available from the Comprehensive R Archive Network (CRAN).
Development versions are available from R-Forge.

Last modified: Jan/26/2013