Development of a machine learning aided system for evaluation of Handwritten Mathematical problems.

Abstract

Assessment/evaluation is an integral part of instruction, as it determines whether or not the goals of education are being met. There is an urgent need to assess student solutions and provision of remarks/results in real time. Evaluation of student solutions is marred by a number of challenges ranging from inappropriate student-to-teacher ratio, social and ethical issues to time and space limitations. This project was undertaken to provide a solution in form of a machine learning aided system that can evaluate student(s)’s handwritten mathematical solutions and provide remark in real time. The system was developed using machine learning models, (Convolutional Neural Network model, CNN) implemented using python programming language. The system interface was developed using Hypertext Markup Language, HTML and JavaScript. The system accuracy is at 78% for the test images sampled. ML aided system have gotten great benefits in education field.