tristanz / OpenIRT (http://people.fas.harvard.edu/~tzajonc/openirt.html)
Bayesian and Maximum Likelihood Estimation of Item Response Theory (IRT) Models
Clone this repository (size: 590.0 KB): HTTPS / SSH
$ hg clone http://bitbucket.org/tristanz/openirt/
| commit 35: | 74f012aa90fc |
| parent 34: | 910cacd4b934 |
| branch: | default |
Got rid of progress bar, didn't work on OSX.
Changed (Δ125 bytes):
raw changeset »
Stata/openirt.sthlp (2 lines added, 2 lines removed)
src/mcmc.h (4 lines added, 3 lines removed)
Up to file-list Stata/openirt.sthlp:
| … | … | @@ -101,8 +101,8 @@ See Das and Zajonc (2009) and Mislevy et |
101 |
101 |
The priors were calibrated using the NAEP item bank and should perform well under a broad range of scenarios.{p_end} |
102 |
102 |
|
103 |
103 |
{pstd}{it:Note on speed}: Estimation can be slow due to the large number of free parameters estimated using MCMC simulation. |
104 |
Users with large data sets may wish to use small subsamples of data before running an analysis on the full sample. |
|
105 |
On many systems a built in progress bar does not currently display in Stata.{p_end} |
|
104 |
Users with large data sets may wish to use small subsamples of data before running an analysis on the full sample. |
|
105 |
{p_end} |
|
106 |
106 |
|
107 |
107 |
{title:General instructions} |
108 |
108 |
| … | … | @@ -1179,12 +1179,13 @@ public: |
1179 |
1179 |
*/ |
1180 |
1180 |
void Iterate(int number_of_iterations, bool progress = false) { |
1181 |
1181 |
if(progress) { |
1182 |
|
|
1182 |
//boost::progress_display show_progress(number_of_iterations); |
|
1183 |
1183 |
for(int iter = 0; iter < number_of_iterations; ++iter) { |
1184 |
|
|
1184 |
cout << "Iteration " << iter << endl; |
|
1185 |
for(int i = 0; i < steps_.size(); ++i) { |
|
1185 |
1186 |
steps_[i].DoStep(); |
1186 |
1187 |
} |
1187 |
|
|
1188 |
//++show_progress; |
|
1188 |
1189 |
} |
1189 |
1190 |
} |
1190 |
1191 |
else { |
