Yelp Rating Prediction with Sentiment and Topic Models.

Public Deposited
Analytics

Downloadable Content

Download PDF
Request Version for Screen Reader
Last Modified
  • February 28, 2019
Creator
  • Liang, Ying
    • School of Information and Library Science
Abstract
  • Online reviews have been widely used for sentiment analysis tasks, for example, sentiment polarity prediction. In this paper, I address the rating prediction problem, using Yelp reviews. A star rating, in most cases, agrees with its review sentiment, which makes sentiment-words a reasonable solution for this task. Topics in reviews, on the other hand, are also likely to influence rating prediction. For example, for a restaurant, a customer may think it has a 5-stars service but the food is just 3-stars. So overall, that customer might give that restaurant a 4-stars rating. Using this idea, in this paper, I investigate whether topics, in addition to sentiment, are helpful in rating prediction task. I incorporated topic model with sentiment model and observed performance improvement.
Date of publication
Subject
DOI
Resource type
Rights statement
  • In Copyright
Advisor
  • Arguello, Jaime
Degree
  • Master of Science in Information Science
Academic concentration
  • Library and Information Science
Degree granting institution
  • University of North Carolina at Chapel Hill
Graduation year
  • 2017
Language
Deposit record
  • 088ad657-f7e9-436d-a0e4-5849d6046222
Date uploaded
  • May 2, 2017

Relations

Parents:

This work has no parents.

Items

Thumbnail Title Date Uploaded Visibility Actions
file details Ying Liang_Master Peper.pdf May 2, 2017 Public Download