Skip to content

Latest commit

 

History

History
33 lines (21 loc) · 1.3 KB

File metadata and controls

33 lines (21 loc) · 1.3 KB

Knowledge Tracing for Online Learning

In a world full of information, data science can help students who don’t have access to personalized learning. Knowledge Tracing means modeling of student knowledge over time. The goal is to accurately predict how students will perform on future interactions. These algorithms can be used to make presonalized online learning possible to every sudent with an internet connection.

Riiid's EdNet data is used for the task. EdNet is the world’s largest open database for AI education containing more than 100 million student interactions. It has data of around 1 million students.

Data

Dataset

Exploratory Data Analysis

Feature Importances

The Feature importance as calculated by LightGBM

This project was made as a part of Riiid! Answer Correctness Prediction, a Code Competition at Kaggle. The task was given by Riiid Labs