An OWA-based hierarchical clustering approach to understanding users’ lifestyles

Based on users’ interactions with social networks, a method to understand users’ life-styles is developed. Descriptions of their lifestyles are obtained from previously reported experiences on these sites. Contextual information and contributed reviews lend insight into which elements are important for different lifestyles. In this paper, an ordered weighted averaging operator (OWA) is integrated with hierarchical clustering in order to find the similarity between users and clusters. Specifically, a two step measure is defined to compare and aggregate two clusters. To illustrate the efficiency of the methodology, a real case is implemented for 499 Yelp reviewers associated with 134,102 reviews across 11 variables and 373 Airbnb reviewers associated with 1,826 reviews across 14 variables.