Automated Exam Paper Marking System for Structured Questions and Block Diagrams

Abstract

Exam paper marking has always been an exhaustive process which requires a lot of time and effort. We intend to address this issue by automating the marking of structured type questions and three diagram type questions, i.e. block diagram, logic circuit, and flowchart questions. A web-based system was developed to collect answers, mark and provide feedback to students. We investigated the application of Natural Language Processing (NLP) when marking structured questions. Graph matching using Depth-First Search (DFS) and Breadth-First Search (BFS) was employed in marking diagram type questions. The accuracy of logic circuit and flowchart diagram marking were tested by simulating and converting the diagrams into Python programs respectively. The implemented system has been made available to users as a plug-in in the opensource learning management system Moodle.