Document Type

Honors Thesis


This paper proposes a novel method to generate ratings from reviews using a Bayesian technique. One of the reasons for the growing trend of online shopping in e-commerce platforms is its transparent review system, where a customer can review and rate a product that becomes open for others to see. Oftentimes, in making a purchase decision, a customer reads these reviews to get feature-specific information about a product. These reviews, however, are becoming increasingly incomprehensible for a person to read in their entirety because of their large volume. Reading a sample of them may create a biased opinion as they do not represent overall reviews. To solve this problem, this project used Bayesian estimation to develop fine-grained, feature-specific ratings of products from the reviews of customers. This task is performed in three steps: (1) mining product features from the reviews of customers (2) identifying the sentiment of the reviews that describe product features (3) generating feature-specific ratings in 5-Point Likert scale. The ratings are generated using the Bayesian approach and are compared with the ones generated using the Frequentist approach.

Publication Date






To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.