This paper tackles Constraint-Based Student Modeling as a new approach to student modelling. This approach splits the subject material into a set of constraints in which any violation to that set will demonstrate a need for that particular area to be revised by the system. This approach simplifies the computational processes required to make it function by utilising pattern matching to determine where the student needs most help.

_{The above diagram displays the interaction data between the student and a constraint-based system. Source: https://www.semanticscholar.org/paper/Applications-of-Data-Mining-in-Constraint-based-Nilakant-Mitrovic/d938ad48fb540b429c4e22100940ed2eac3fd112/figure/5}

This technique can be utilised online to view the ongoing performance of the students as well as offline to analyse the results produced from the work they have completed. For the purposes of this paper testing was conducted solely for the offline approach.

Testing was conducted by choosing a type of task that the student must work out. The constraints were then established, determining the types of actions that the student must peform in order to fulfill that task correctly. Several test cases where then created to display how the system would approach being given sets of steps and the answers reached as a conclusion, after which the system would check for any constraint violations and then determine the best approach to finding the correct solution.

_{The above is the correct solution path generated by the system which it provides to users whenever the syste detects a constraint violation in the user's approach to the problem.}

Due to the theoretical nature of this model the researchers determined that this system does not provide any realistic or beneficial student models. They instead wanted to encourage future research into the potential of this learning technique.