Ryan C. Sanchez

In some instances, the initial guess for the multi-dimensional estimates is not SPD. This error is because the polychoric correlation contingency table is light on samples. Add a check and...

The Readme file is getting a little long in the tooth. Add some drop-down menus to make it a little easier to read.

I'm not sure if the current implementation will work at scale? I need to do an investigation at the performance. My initial thought is underflow might occur for mml due...

I treat Rasch and 1PL as nested, which means the mml Rasch model assumes a standardized normal population. This is a non-standard way of doing things

Rasch with ability guessing

1

Implement the rasch model with ability guessing. This was suggested as alternative to the random guessing usually assumed.

Python 3.10

1

Check compatibility with python 3.10 and setup CI/CD

Current parameter estimation techniques assume order invariance. Investigate discrete state-space parameter recovery of the form θn = A θ n-1 + B μ m = C θn + D μ...

Issues #76 added polytomous ability estimation as an output to most mml methods, would still like a stand-alone function similar for dichotomous methods.

Continue webpage documentation on "how-to use"

The graded unfolding model is an ideal point IRT model and an extension of the PCM #9. The framework for this model reuses a lot from the partial credit model...