Abstract:
Not responding to consumer needs in the hotel business leads to decreased satisfaction, negative word-of-mouth, and inefficient resource allocation. Machine learning can be applied to analyzing and predicting guests needs and feelings to improve hotel services. This research aimed to (1) improve and assess the performance of using machine learning techniques for perceiving tourists hotel staying behaviors and (2) develop an application for connecting the predictive model to hotel businesses. The dataset was collected from 300 tourists opinions by Google Form. The data imputation used the K-Nearest Neighbors Algorithm--KNN, and some effective features were selected for training by information gain. The selected features were sent to train and test performance by machine learning algorithms including Decision Trees--DT, K-Nearest Neighbors Algorithm--KNN, Neural Networks--NN, and Support Vector Machines--SVM. This research found that the best-performing machine learning algorithm for predicting tourists repeated hotel stays and hotel recommendations for other people was KNN. The KNN with k=3, k=5, and k=7 can give higher prediction accuracy than other algorithms at 98.67%. The KNN was implemented to serve hotel businesses via the friendly web application.