Sentiment Analysis using Token2Vec and LSTMs : User Review Analyzing Module
Achira Jeewaka Shamal, Rankothge Gishan Hiranya Pemathilake, Sachith Paramie Karunathilake, Gamage Upeksha Ganegoda · 2018
Sentiment analysis can be defined as the process of identifying the speaker's attitude towards a particular subject. Sentiment analysis is very useful for businesses when it comes to monitoring and analyzing user reviews of their products. Because, it is capable of automatically classifying the overall opinion of customers of a specific product or a brand by analyzing the feedback that they have already given. Furthermore, companies can easily collect relevant user reviews from social media platforms such as Facebook, Twitter, blogger etc. Because of those reasons sentiment analysis has become a popular topic in natural language processing research areas. This paper proposes a Long Short-Term Memory based deep learning approach to perform sentiment analysis of social media user reviews on electronics domain. Specially this model is capable of analyzing the sentiment of reviews which consists emojis and social acronyms.