Predicting the outcome of purchasing a hotel package to assist policymakers using machine learning

Abstract

Given the growing and massive shock to the hotel industry, the coronavirus (COVID- 19) outbreak has generated an unexpected problem. The hotel industry is currently facing a big obstacle during the COVID period, where strategies designed to promote sustainability will play an essential role for the industry. Hotel policymakers would need to learn from the COVID-19 issue in order to increase sales, improve crisis management policies, and better prepare targets and the industry as a whole to respond to unforeseen situations. It would be preferable if the hotel sector used information and technology and focused on automated ways of learning and forecasting from historical data. The objective of this research is to study customer information in order to make a recommendation to the policy maker and marketing team, as well as to develop a model to predict who will buy the newly launched vacation package. It will assist the hotel industry in enabling and establishing a viable business model for growing their customer base. As a result, we’re adopting machine learning to predict which customers will be interested in purchasing the hotel package. A survey was used to gather data for this study. The data was evaluated to uncover key elements for our study, and we employed seven algorithms, namely, Random Forest Classifier, K Neighbors Classifier, Naive Bayes, AdaBoost, Support Vector Machine, Logistic Regression, and Gradient Boosting algorithm, to predict potential customers based on those features. We have obtained an accuracy of 95% while also reducing the number of false negatives by using the Gradient Boosting Classifier, Support Vector Machine, Random Forest Classifier, Naive Bayes, Logistic Regression. Our findings show that, without a team of analysts, our analysis and suggested approach can provide the finest insight to policymakers to help them make better decisions.