Sentiment analysis from unstructured hotel reviews data in social network using deep learning techniques

The Internet is benefiting the people to access and provide information. One of the platforms to share the information is online social networks such as blogs, Facebook, and Twitter. People express their sentiments on the products or services used through reviews. These reviews are more important for business organisations to grow in the market. But the review information provided by people contains sentences with spelling mistakes, abbreviations, symbols, emoticons and special characters. It is complex and consumes more time to understand whether it is a good or bad or average sentiment of the people. In this work, fine tuned Bidirectional Encoder Representations from Transformers (BERT) and Natural Language Processing (NLP) Techniques are proposed to analyse the hotel reviews information for customer sentiment analysis and Recurrent Neural Network (RNN) also explored. Comparative study is carried out to understand the performance of the proposed model.