Automated Assessment System Using Machine Learning Libraries

Abstract

Assessment and the grading of students is a task that has been done for as long as school has existed. This was previously done by teachers in primary and secondary, lecturers for institutions like JAMB and lecturers in schools. Up until now students’ marks were influenced by other external factors such as bad handwriting, lengthy paragraphs, roundabout way of speaking rather than going straight to the point and the sheer number of assignments the lecturer has to mark. This has resulted in students getting lower or higher marks than they should be awarded. This project is to create an ML (Machine Learning) powered assessment system that will take the assignment questions and the marking scheme and award the student the marks similar to what the ideal lecturer would have given. This will also reduce the time the lecturers spend on marking and ensure the students get their results on time. This project will be made with Python and machine learning and will be tested with a number of potential answers to the questions and their grading’s. This will enable system to be able to grade assignments as soon as they are uploaded. This research will be limited by the fact that the system can only handle the marking of short sentences accurately and not long paragraphs. The system is also limited by the fact that it can only mark with the aid of the marking scheme and not without it so it is not a truly intelligent model in that regard. The research showed that the system is indeed capable of obtaining the similarity between two paragraphed answers provided but it needs extras to produce the most accurate results.