ML-Correctness-prediction

This is the CSC311 Final Project in University of Toronto. Group member: Zuoyu Wang, Siqi Chen.

We worked in pairs and build machine learning algorithms to predict whether a student can correctly answer a specific diagnostic question based on the student’s previous answers to other questions and other students’ responses.

Algorithms involved:

1. k-Nearest Neighbor using scikit-learn.
2. Item Response Theory.
3. Neural Networks using PyTorch.
4. Bagging ensemble
5. Adaboost using NN as base classifier.

• Data is not shown in this repository to protect copyrights.