Clustering Hostels Data for Customer Preferences using K-Prototype Algorithm

Abstract

Reviews on hostel booking platforms can be used to collect rating data from customers which can be useful for service providers and customers in the future to determine hostel choices based on their individual preferences. The rating data can be used to determine the pattern of hostel selection by customers and can provide suggestions according to customer preferences in choosing hostels in certain areas. In this study, the data will be grouped using the k-prototype algorithm which is a combination of k-means and k-mode algorithms so that it is possible to group mixed attributes. The results of this study are to determine the data segment in accordance with user behavior in selecting the hostel at the time, location of the hostel and certain conditions, so that by knowing the segmentation profile, the hostel service provider can easily provide product promotions to segmented customers.