1 / 13

Predicting Academic Performance (Learning Dynamic Bayesian Networks)

Predicting Academic Performance (Learning Dynamic Bayesian Networks). Mashael Al- Luhaybi Ph.D student in Machine Learning Principal Supervisor: Dr Allan Tucker. Second Supervisor: Dr Stephen Swift. Presentation Outline. 1. Aim & Objectives. 2. Methodology. 3. Experimental Results.

vshirley
Download Presentation

Predicting Academic Performance (Learning Dynamic Bayesian Networks)

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Predicting Academic Performance(Learning Dynamic Bayesian Networks) • MashaelAl-Luhaybi • Ph.D student in Machine Learning • Principal Supervisor: Dr Allan Tucker Second Supervisor: Dr Stephen Swift

  2. Presentation Outline 1 Aim & Objectives 2 Methodology 3 Experimental Results 4 Future Work

  3. Data-driven AI?

  4. Statistics * The Higher Education Statistics Agency (HESA) in the UK 26,000 788,355 (2012/13) 757,300 (2016/17) HE Qualifications Drop out (England) 2015

  5. Aim & Objectives • The essential aim of this research is to predict the academic performance of university students using probabilistic modelling taking into consideration the bias issue of educational datasets. • Objectives: • Identify the most predictive features (attributes) that affecting the prediction of students academic performance in higher education institutions. • Develop classification models (BNs) that predict students overall performance based on time series data & other extracted features from Blackboard learn. • Find solutions for imbalanced educational datasets.

  6. Methodology Proposed Approach

  7. Methodology Proposed Approach

  8. Experimental Results 1- Temporal Profiling Online Trajectories (DTW + Hierarchal Clustering)

  9. Experimental Results 2- Learning Bayesian Networks (Bootstrapped Educational data)

  10. Experimental Results Learning Bayesian Networks (Bootstrapped Educational data)

  11. Experimental Results Learning Bayesian Networks (Bootstrapped Educational data)

  12. Future Work • Explore the extension of these Bayesian models with the investigation of latent attributes, with the aim of capture hidden factors that may influence the dynamics of students academic performance. • Compare our approach with other resampling approaches for time-series data (investigating Grade-based bootstrapping). • Explore some methods to solve bias issues of time-series dataset.

More Related