Analyzing Restaurant Reviews to Predict Customer Satisfaction Trends

Abstract

Numerous studies comparing the quality of various restaurants' cuisines have been conducted. Area, average cost for two people, votes, cuisines, mostly flavour, and restaurant type are some of the criteria used to evaluate restaurants. Finding out which restaurants people like to eat at and reading reviews of such places is the primary objective here. The goal of this study is to develop a model that can foretell if a review of the eatery will be favourable or negative. A number of prediction algorithms, including Multinomial Naive Bayes, SVC, XGB Regressor, Pipeline, and Logistic Regression, will be utilised to achieve this goal. Finally, we'd like to identify the "best" model that can forecast the reviewer's emotional tone.