Rule-based XGBoost model with power transformation effectively detects incentivised hospitality reviews, offering scalable, generalisable protection for platform credibility.
Authors
Syed Abdullah Ashraf, Assistant Professor, Jindal Global Business School Department of Information Systems & Analytics, O.P. Jindal Global University, Haryana, Sonipat, Haryana, India
Aariz Faizan Javed, Indian Institute of Management Ranchi, 5th Floor, Suchana Bhawan, Meur’s Road, Audrey House Campus, Jharkhand, Ranchi, 834008, India
Pradip Kumar Bala, Indian Institute of Management Ranchi, 5th Floor, Suchana Bhawan, Meur’s Road, Audrey House Campus, Jharkhand, Ranchi, 834008, India
Rashmi Jain, Department of Information Management and Business Analytics, Feliciano School of Business, Montclair State University, 1 Normal Ave, Montclair, 07043, NJ, United States
Zainab Fatma, Department of English, Aligarh Muslim University, Aligarh, 202002, India
Summary
Incentivised reviews are a permanent threat to the credibility of information available on a platform. They not only interfere with the consumer decision-making process but also impact the market dynamics. We have proposed a rule-based method for identifying incentivised reviews in the hospitality domain. We then extracted several features from the meta-feature. These features were transformed using mathematical functions. The study showed that power transformation along with XGBoost is best suited for the task. Our work has both practical and managerial implications. Our model is lightweight, scalable, and generalisable. Moreover, platforms can use our model with a fake review detection method to safeguard the interest of honest, hardworking sellers and buyers looking for trustworthy information. Based on our research, our work is among the first few to address incentivised review detection in the hospitality sector.
Published in: International Journal of Business Information Systems
To read the full article, please click here.
