Automation of Non-Classroom Courses using Machine Learning Techniques

Abstract

Non-classroom courses in higher education significantly encourage creative thinking. Awarding credits for accomplishments through non-classroom courses encourage students for more participation and improve their problem-solving skills. Machine learning techniques help to automate the maintenance of such non-classroom courses. This paper proposes a process to introduce and automate such a non-classroom. Firstly, students will submit a softcopy of the certificates/proofs of their accomplishments such as winning/participating in a competition/event through a web-based user interface. Secondly, the authenticity of these proofs/certificates is tested using text mining, image processing, and web scraping techniques. Text is extracted from the submitted certificates using Tesseract Optical Character Recognition (OCR) Engine and stored as a MySQL database which is matched with the information provided by the student in the web interface and other sources such as flyer of the event. Further, the extracted text is used as search criteria for web scraping and matched with the scraped text. Further, the authentication test can be performed with these keywords and tags by search through social media. Facial Encodings and Recognition is done on the images submitted to authorize the user. Finally, the evaluation is performed and the final result indicating the credits/grade is revealed to the user.