The repository contains data and scripts for data visualisation and simulation used in the EDM'23 paper "Fast Dynamic Difficulty Adjustment for Intelligent Tutoring Systems with Small Datasets".
To run all the scripts here, you need Python 3, Node.js, and MongoDB installed. The required Python packages are listed in requirements.txt.
The sections below list the different sub-applications that are contained in the repository. Each of these can be easily launched as described.
The web application for the user study is located in webapp. You first need to install packages by opening the command line, entering the directory webapp, and typing npm install. Then you can start the application with node index.js, and then the application is accessible at localhost/index.html. You can change the method of exercise selection by changing the setting in the header of webapp/index.js.
To run simulations, first you need to start the web application as described above, since the Python script calls the API hosted in the web application. Then run the Python script simulation/simulate_user.py. The graphic results from the simulations are saved in plots_simulation.
The collected data is located in data, containing the data used for analysis, for training the DDA model, and the intermediate results from the knowledge tracing model used to perform further analysis. These are used by scripts in data-analysis to generate various plots and perform statistical tests. The plots are stored in plots.