Abstract: Nowadays, due to users critically reviewing the restaurant and related services, there is an abundance of the reviews that are available on the internet. There is an opportunity to get more detailed information from the customer’s review using sentiment analysis. Certain steps are taken to conduct this sentiment analysis. Implementing a machine learning based approach for sentiment analysis increases the performance and accuracy of the sentiment analysis task, as compared to the lexicon based approach. The accuracy achieved for the chosen dataset was between 75% to 80% and it can be increased furthermore with some changes to the process. The model is trained using domain specific data, so it performs better than a lexicon based approach.
Keywords: Sentiment analysis, Data mining, ML, NLP, SVM
| DOI: 10.17148/IJARCCE.2021.10545