Automated Grading and Feedback of Programming Assignments

Currently, multiple graders are required to grade and provide feedback to many students.
As a result, there is a discrepancy between the grades awarded, and the quality of feedback received, partly due to increased student numbers.

This project aims to create a comprehensive suite of tools for auto-grading and generating meaningful feedback for novice Java programmers, focusing on maintainability, readability and documentation.
Auto-grading and auto-feedback help to reduce the variance in awarded grades, when marked by multiple graders, and provide more consistent high-quality feedback.

A mix of metrics and machine learning will be used to complete the project's auto-grading portion. In particular, existing programming metrics and metrics from other areas, such as literature, will be evaluated to see if they aid in the marking process.
Generating feedback will use natural language processing tools and machine learning to generate feedback, which will then be evaluated to see how meaningful it is in a given context.