Sentiment Analysis and Visualization of Reviews for Healthcare Service Providers Using Naïve Bayes

• Published in: AIP Conference Proceedings
• Main Author: Saman F.I.; Chua N.F.C.J.; Jamrus F.N.
• Format: Conference paper
• Language: English
• Published: American Institute of Physics 2024
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85191975884&doi=10.1063%2f5.0181989&partnerID=40&md5=a8b46b3d0bf5cb454b3db93d0c9b7bca

Improving and maintaining the hospital service quality is considered a global concern. Healthcare service providers widely use online patient feedback to measure and improve the quality of care in healthcare services from the patients' perspectives. Similarly, customers will read reviews before deciding on which hospital to receive their treatment making reviews to be useful for both customers and hospital providers. However, due to the unstructured nature of user reviews, it has been challenging to extract useful information where the number of reviews could range from hundreds to thousands with varied opinions thus making it difficult to extract and analyze the data. Sentiment analysis is widely recognized as one of the effective approaches for analyzing the sentiments of data in terms of people's opinion. This paper describes the implementation of sentiment analysis on patients' reviews from six hospitals in Kuala Lumpur using a Naïve Bayes classifier and the results are presented in the form of a visualization dashboard to let patients and hospitals understand public opinion on the hospital services. Top occurring words associated with hospitals have been identified and it can provide better insights on service quality of the providers. For future works it is suggested to use larger amount of labeled data to improve performance and use different classifiers for performance comparison. © 2024 American Institute of Physics Inc.. All rights reserved.