Plan

boldface text is still to be completed. Other text means it has been written/revised in some way though not necessarily complete.

Introduction

• Briefly explain the general structure of the report, discuss the final outcomes

Background

• Infandango and current feedback. Also talk about the course and gathered data
• Machine learning to solve problems
• Literature
  • Khan blog
  • Programming assessment
  • Edward Tufte summary
  • some machine learning paper about optimisation

Design

• Language and tools
• Proposed Design
  • Machine learning model
  • Visualisation
• How it will integrate with Infandango

Choosing a model

• Feature selection
  • do we retain question identity? Yes and no Preliminary models
  • Classifiers }
  • Regression } For both of these explain the libraries and model details
• Training and optimising
  • K-fold cross validation
• Results
• Conclusion

Implementation

• Overview
• CSS Design
• Prediction Model
• System Integration
• Improvements

Conclusion

• Stress the main acheivements:
  • a model that predicts the score
  • give the user more feedback, and it has been integrated and is working
• What could be done
  • Use different features
  • Tests