Machine learning algorithms as tools for student success prediction
summary
Applies supervised ML (decision trees, k‑NN, neural nets, Bayesian) to 400 students across three generations combining Moodle logs, surveys and YouTube analytics; follows CRISP‑DM, runs sensitivity analysis and predictor ranking. Finds high school GPA and entrance exam as top predictors, with first-year grade and LMS activity also strong; female students show higher LMS activity and grades. Video duration and topic complexity negatively correlate with views (10–15 minute clips perform best). Predictors shift only when the education system changes.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 0:46Context
- 2:02Brief Context
- 2:32Machine Learning Algorithms
- 3:12Framework
- 4:44Research Papers
- 4:59Literature Review
- 6:55Data Source
- 8:16Results
- 8:42First Grade
- 9:55Statement
- 10:35Data
- 11:03Graphs
- 14:06YouTube Analytics
- 17:12Sensitivity Analysis Results
- 18:29Correlation Analysis
- 19:48Conclusion
- 20:29Future work
- 20:56Research questions
- 21:54Questions
related talks
19mAdvancing Personalized Learning Analytics with AI and EdTech
03mApplying data and machine learning to scale education
22mPredict the future (or don't) with machine learning in applications
22mMachine learning in Education
39mForecasting With Classical and Machine Learning Methods
15mOperationalizing Predictive Analytics in Education
52mMachine Learning Crash Course
27mData Science and ML with an emphasis on Banking...