Analysis of Public Sentiment on Dhaka Metro Rail with Transformer Based Architectures

Dhaka Metro Rail is a mesmerizing initiative of the Bangladesh Government that has brought comfort to people. This project got a huge appreciation from people for its continuous service. Mass people use the train for rambling, going to the workplace, etc. Again, for some unwanted circumstances, Dhaka Metro Rail gleaned negative criticism from various public.In this research, we analyzed public sentiment on Dhaka Metro Rail from popular social media using recent NLP architectures like BERT, XLNet and Electra, where BERT outperformed than other with an accuracy of 91.11% in correctly identifying the sentiments of the public comments.The sentiment of a comment is identified into predefined categories, such as positive, negative, or neutral.