Deep Learning Approach for Sentimental Analysis of Hotel Review on Bengali text

We live in an age of technology. Technology is advancing day by day. People in India (Kolkata), Bangladesh, who use the Bengali language as a communication language, use the internet daily. Internet users are increasing day by day. Moreover, people like to visit many tourist places. So, they have to stay in a hotel or resort. They give various reviews on websites and many social media. But not everybody gives thoughts in the English language. People in Bangladesh, India (Kolkata) give reviews in Bengali. So, this is hard to detect for the hotel management system. So, in this paper, we are planning to build a model that can analyze the Bengali text and detect whether the reviews are bad or good since there is no work on Bengali text in this particular area. The datas are collected by surveys, various social media reviews, ratings, etc., and then label those data. After cleaning and extracting multiple features, the datas were trained into DL(Deep Learning) and ML(Machine Learning) models. At the end, it is found that the Long Term Short Term (LSTM) comparatively gives better results than other models. This work will create a significant effect on the hotel and tourism industries.