Item Selection for Parallel Test Construction
2024/05This project uses genetic algorithms to assist in item selection for parallel test construction. With one click, users can generate two exam papers of equivalent difficulty with completely different questions. This is especially useful for teachers needing two versions of a test, such as one for make-up exams, ensuring fairness in difficulty and avoiding question repetition and exposure.
Technologies
Introduction
"Item Selection for Parallel Test Construction" is a project I completed for the AI course in the second semester of 2024. The frontend was built using React JS, React Query, and the MUI framework. The backend was developed with Node JS and Express JS, and MongoDB was used as the NoSQL database. Both frontend and backend were deployed on the Zeabur platform.
In addition, to apply theories of Genetic Algorithms and Parallel Test Construction in practice, I studied relevant research papers and implemented the concepts in the backend API, enabling users to perform item selection through the frontend interface.
Project Links
Key Features
- Add new items to the question bank
- View all items in the question bank
- Automatically generate test items based on parallel test theory
- Generate items using genetic algorithm
- Download the test paper with one click
- Each test item displays its difficulty