A Novel Perspective to Mining Online Hotel Reviews Based on Biterm Topic Model

Nowadays, the rapidly increasing number of online hotel reviews have played a significant role in customers' satisfaction attributes mining and businesses' customer relationship management. Especially, for global hotel chains, they provide services for customers from different countries. Yet different customers have different service quality expectation based on their culture. However, very little is known about what are the main service quality attributes they care about respectively. Thus, in this paper, Europe and China, two representative markets in the tourism industry, are selected to conduct research. An experimental study is conducted and a sample of 515,738 reviews from booking.com and 10,000 reviews from ctrip.com are used. Besides, this study selects a biterm topic model that is suitable to deal with short review texts to mine the main attributes. Finally, seven categories of attributes are found. Results demonstrate that European customers care more about environment, facility, price, and food, while Chinese customers pay more attention to cleanness. The study enriches the stream of research on hotel reviews in the hospitality industry from an innovative perspective and provides useful implications for global hotel chain managers to meet their guests' satisfaction and promote their service qualities.