Myanmar Lexicon Based Sentiment Analysis on Hotel Reviews

Abstract

As social media and digital communication use increases in Myanmar, sentiment analysis is being used more and more in business, politics, and social trends. Big social data analytics is a valuable tool that can be utilized to uncover significant information from social user data. This methodology integrates diverse statistical techniques, sentiment analysis, multimedia administration, and social media analytics to anticipate and predict individuals and examine patterns. Natural Language Processing (NLP) tools and frameworks are becoming more customizable and easily accessible, which makes the process of creating language models unique to Myanmar easier. The proposed system's lexicon will have six categories of aspects (Room, Staff, Facilities, Location, Value, General), together with their corresponding subcategories and opinion terms. After that, word2vec is used to train the reviews of the annotated corpus and create a word embedding model. Because of the nature of the Myanmar language, it is particularly more difficult to perform aspect-level opinion mining on reviews about Myanmar. As a result, the proposed system's primary goal is to employ syntactic patterns and rules to extract pertinent pairs of attributes and opinion terms from user evaluations. The proposed method could be increase the accuracy of sentiment analysis on social media postings written in Myanmar.