Recent Changes - Search:



Bayesian Networks in Educational Assessment

Cognition and Assessment


edit SideBar

BN /


5 item IRT

These are a simple 5-item IRT model built with and without a context effect used for illustrating Bayesian networks in the tutorial. These nets also appear in Chapter 6 of Bayesian Networks in Educational Assessment.

All of these examples together can be found at tarball (tarball,zip)

5 Item IRT

This example is a simple 5-item IRT example. The proficiency variable, theta, is discritized into 5 levels and given a triangular distribution (which is close to normal). The conditional probabilities are created by evaluating the Rasch model at the 5 fixed values of theta. The difficulty parameters is increasing from Item 1 to Item 5.

Picture of IRT 5 example.

Netica Network

A variant network for use with EM learning is also available (Duanli, can you edit this, as I'm not sure what this one is for). Netica Network

5 Item IRT with context effect

This example adds a context effect? between Items 3 and 4. Marginal probabilities are adjusted so that they match those of the previous example.

Picture of IRT 5 with context effect example.

Netica Network

Dibello--Samejima distribution illustration

This is a simple illustration of the DiBello--Samejima? model for partial credit Invalid BibTex Entry! items. It is also an illustration of Netica?'s equation feature. The free floating nodes in the model are the difficulty? (B) and discrimination? parameters. You can change those and rebuild the tables to see the effects of those parameters.

DiBello--Samejima example with 3-level observable.

DiBello-Samejima model with 3-level observable.

Netica Network

DiBello--Samejima example with 4-level observable.

DiBello-Samejima model with 3-level observable.

Netica Network

Edit - History - Print - Recent Changes - Search
Page last modified on March 19, 2015, at 03:43 PM