A Novel Approach to Estimation Precision and Recall for Star Rating Online Customers Based on Negative Hotel Reviews using Multinomial Naive Bayes over Multischeme Classifier

To estimate precision and recall for hotel star rating using sentiment content. Majority class classifier with sample size (N=10) and Multinomial Naive Bayes with sample size (N=10) were iterated at different times for predicting accuracy percentage of hotel review. The F1 measure used in prediction to probabilities which helps to improve the prediction of accuracy percentage. The sigmoid function used in Simple majority classifier prediction to probability which helps to improve the prediction of accuracy. There was a statistical significance between Multinomial Naive Bayes and Majority class classifiers (p=0.00). Results proved that Multinomial Naive Bayes got significant results with 68% accuracy compared to Majority Class Classifier with 67% accuracy. Multinomial Naive Bayes is a simple and most effective algorithm to build fast machine learning models. Multinomial Naive bayes with f1 measure helps in predicting with more accuracy percentage of hotel review.