ORCID Identifier(s)

0000-0002-4331-7759

Graduation Semester and Year

2020

Language

English

Document Type

Thesis

Degree Name

Master of Science in Computer Science

Department

Computer Science and Engineering

First Advisor

Deokgun Park

Abstract

Evaluating and providing feedback to hundreds of free text assignments in an online environment is a challenging task for an instructor where he has to scan through essays to identify perspectives that are expected to appear in those essays. Reading large number of essays and then finding themes and providing customized feedback are time-consuming process. We have proposed a text analytics system named EssayIQ that aids course instructor in identifying assignment themes, providing theme presence statistics and giving feedback to learners. To the best of our knowledge, this is the first system that analyzes free text assignments in line with the instructor defined themes. Through our experiments on one online course, we have shown that model based on sentence-level semantic embedding outperforms word and phrase based embedding models. We have also shown that EssayIQ system can identify themes and can generate overall theme statistics for over hundred submissions within minutes with minimal theme knowledge intake by EssayIQ. The theme identification quality of EssayIQ system is also comparable with human/coach annotators. The code of this project is publicly available in github (https://github.com/Shadek07/EssayIQ).

Keywords

Automated concept identification, Visual analytics, Yext analytics, Phrase2vec, Online learning, Word2vec, Universal sentence encoder, Reflection essay, Debriefing essay

Disciplines

Computer Sciences | Physical Sciences and Mathematics

Comments

Degree granted by The University of Texas at Arlington

Download

Share

COinS