KST-toolbox

The first MATLAB toolbox entirely devoted to deterministic and probabilistic knowledge structure theory.

KST toolbox

Toolbox authors: Andrea Brancaccio, Luca Stefanutti & Debora de Chiusole

Other toolbox references: Brancaccio A., de Chiusole D., Wickelmaier F. (2024). Software. In Heller J. & Stefanutti L. Knowledge structures Recent developments in Theory and Application. Vol. 7. World Scientific. DOI

The first MATLAB toolbox entirely devoted to knowledge structure theory. Look at GettingStarted.mlx for more information on how to use the Toolbox.

Data Structure

Data Structure	Description
states	is a binary matrix k times n where k is number of knowledge states and n the number of problems in the domain Q. Each row 0 < i <= k is a vector representation of a knowledge state where a cell (i,j)= 1 if problem j belongs to the knowledge states, 0 otherwise.
data	is a binary matrix s times n where s is number number of participant and n the number of problems in the domain Q. Each row 0 < i <= s is a vector representation of participants i response pattern. A cell (i,j)= 1 means that participant i responded correctly at problem j, 0 otherwise.
model	is the main output from all functions that applied a version of the basic local Independence model. It is a structure with multiple fields corresponding to all the relevant features of the fitted models such as the parameters, absolute fit statistics, and model comparison statistics.

Building and plotting knowledge structures

Function	Description
skillmap	Delineate the knowledge structure from a skill map
base2struct	Build a knowledge space from its basis (algorithm described in Doignon & Falmagne, 1999, p. 32)
plot_knowledge_structure	Visualize the Hasse diagram of a knowledge structure

Fitting the Basic Local Indipendence Model

Function	Description
blim	Estimate the BLIM by maximum-likelihood via the EM algorithm
sample	Sample based on the BLIM’s parameters
bootblim	Compute the Chi-square p-value of the BLIM via parametric bootstrap
blimfit	Estimate the BLIM and compute its p-value via bootstrap

Testing the Local Identifiability

Function	Description
blimit	basic local independence model identification test see Stefanutti, L., Heller, J., Anselmi, P., and Robusto, E. (2012).

KST-toolbox is licensed under CC BY-NC-SA 4.0