Sentiment Analysis of Restaurant Reviews in Social Media using Naïve Bayes

Abstract

There are currently huge reviews that have increased every day of customer actions in online media, such as Twitter about the restaurant industry, which needs more focus for customers by constantly enhancing customer service. Satisfaction of the customer is a gateway to customer service. Sentiment analysis is a way of defining negative or positive opinions when classifying documents. It was almost impossible to evaluate the data of reviews in social media manually. There are various data mining algorithms used for evaluating and analysis. This paper analyzes the sentiment of users' tweets on the subjects of restaurants by using the sentiment classification algorithms of Naïve Bayes (NB). The system suggested is capable of performing sentiments analysis of tweets. For this reason, a sample of real data was collected from Twitter for customer reviews. A python programming language has been used for implementing the proposed system. The results are similar to the fact that customers are highly positive about restaurants and people, primarily people who are positive about restaurants. The results are measured using common measures for performance assessment, including precision, recall, accuracy and error rate. They were obtained as 68%, 80.07%, 73% and 27% respectively.

Keywords

• customer satisfaction
• naïve bayes
• sentiment analysis