Understanding Homework Reviews Through Sentiment Classification

Author

Zachary Mekus, Washington University in St. LouisFollow

Date of Award

Spring 4-22-2019

Author's Department

Computer Science & Engineering

Degree Name

Master of Science (MS)

Degree Type

Thesis

Abstract

This thesis uses a naive bayes sentiment classifier to analyze six semesters of homework review data from CSE427S. Experiments describe the benefits of an automated classification system and explore original ways of reducing the number of features and reviews. A new algorithm is proposed that tries to take advantage of aspects of the review data that limit classification accuracy. This analysis can be used to help guide the process of automatically using short reviews to understand student sentiment.

Language

English (en)

Chair

Marion Neumann

Committee Members

Marion Neumann Yevgeniy Vorobeychik Ron Cytron

Comments

Permanent URL: https://doi.org/7936/40p5-c027