assessing-knowledge-state

Adaptive assessment diagnoses a student's knowledge state by iteratively selecting fringe-based items from a knowledge graph and updating posterior probabilities using BLIM, PoLIM, or MOCLIM to infer competence.

BLIM, PoLIM, and MOCLIM are probabilistic models used for posterior updates during adaptive assessment, each offering different mathematical approaches to infer a student's knowledge and competence states from their responses.

You can run an ALEKS-style assessment by loading a knowledge-space graph in JSON format, then using built-in utilities to enumerate states, select fringe-based items, and perform posterior updates for a specific student.

Yes, the tool reads and writes knowledge-space graphs stored as JSON files in the graphs directory, directly integrating your existing knowledge-space structures into the adaptive assessment pipeline.

The assessment pipeline provides utility functions for state enumeration, posterior updates, session analytics, and state validation to track student progress and ensure assessment accuracy throughout the session.

Fringe-based item selection is used during adaptive assessment to target questions at the boundary of a student's current knowledge state, efficiently narrowing down competence inference with each response.