edm-learning-analytics

What problem does it solve?

Educational data mining turns raw LMS interaction logs and assessment results into actionable predictions about learning progress, dropout risk, and behavioral patterns.

Core Features & Use Cases

• Model mastery with Bayesian Knowledge Tracing (BKT): estimate per-student probability of knowing each skill over time from attempt-level correctness.
• Forecast dropout with Cox survival analysis: predict time-to-dropout and quantify risk using hazard ratios and survival curves.
• Explain learning behavior with learning curves, sequence mining, and IRT: fit power-law learning curves, discover frequent activity sequences (PrefixSpan), and estimate student abilities/item difficulties with a 1PL Rasch IRT model.
• Use Case: For a course exported from Moodle/Canvas, use early engagement and quiz outcomes to identify students likely to struggle or drop out, and reveal which learning sequences correlate with success.

Quick Start

Use the skill to ingest your Moodle/Canvas CSV logs and run BKT, Cox dropout prediction, and sequence mining to produce mastery and at-risk insights for each student.