A hybrid machine learning approach to hotel sales rank prediction

One of the challenges that the hospitality and tourism industry faces is determining the best-rated and ideal hotels for people with customized preferences. Users belong to various demographic groups, and the factors they consider when selecting a hotel depend on their priorities at the time. Therefore, to provide appropriate recommendations tailored to the individual preferences of users, forecasting customer demand is required, for which hotel sales rank prediction models are to be developed. In this regard, the present paper aims to develop a customized hotel recommendation model for sales rank prediction that considers factors like distance from a strategic location, online user ratings, word-of-mouth rating, hotel tariff, and customer reviews, using the aggregated data set of Indian hotels from trivago.com. Results show that the Artificial Neural Network algorithm predicts sales rank better than the Random Forest and Gradient Boosting algorithms. Implications for practice are provided.

Automated summary

This study aims to develop a customized hotel recommendation model for predicting sales rank. By considering factors such as distance from a strategic location, online ratings, word-of-mouth rating, hotel tariff, and customer reviews, the Artificial Neural Network algorithm can predict sales rank more accurately.