An Application of Ordered Weighted Averaging Operators to Customer Classification in Hotels

An algorithm widely used in hotel companies for demand analysis is the so-called K-means. The aforementioned based on use of Euclidean distance as a dissimilarity measure and this fact can cause main handicap. Concretely, provides global difference between values descriptive variables that blur relative differences each component separately and, hence, cluster assign custom to an incorrect cluster. In order avoid drawback, paper proposes application Ordered Weighted Averaging (OWA) operators OWA-based K-means clustering customers staying at real five-star hotel, located mature sun-and-beach area, according their propensity spend. It must be pointed out calculates distances it sensitive separately. All experiments show OWA operator improves performance classical up 21.6% reduces number convergence iterations 48.46%. Such improvement has been tested through ground truth, designed by marketing department firm, which states tourist belongs. Moreover, customer classification achieved regardless season stays hotel. these facts confirm could appropriate tool classifying tourists purely exploratory predictive stages. Furthermore, new methodology implemented without requiring radical changes implementation data processing crucial so incorporated into control panel additional costs.