Diagnostic Test for Inferring Learning Errors

This system uses Bayesian probability to diagnose students' learning errors. Teachers can add questions and define error types. After students take the test, the system instantly calculates their familiarity with each error type and predicts their correctness probability. At the end of the test, unfamiliar error types are ranked, and learning videos provided by the teacher are recommended for review.

Technologies

Introduction

"Diagnosing Learning Errors through Diagnostic Tests" is a project I completed for the AI course in the second semester of 2024. The frontend was built using React JS and the MUI framework. The backend was developed with Node JS and Express JS, and MongoDB was used as the NoSQL database. Both the frontend and backend were deployed on the Zeabur platform.

To apply the theory of Bayesian probability in practice, I studied relevant research papers and implemented the concepts into the backend API. This allows users to take diagnostic tests through the frontend interface, view real-time correct and incorrect rates, understand their recognition level of different question types, and review their error types after the test to identify weaker areas.