ABSTRACT

Recommender systems are progressing as a vital part of every industry with no exemption to travel and tourism segment. Considering the exponential proliferation in social media usage and huge volume of data being spawned through this channel, it can be considered as a vital source of input data for modern recommender systems. This in turn resulted in the need of efficient and effective mechanisms for contextualized information retrieval. Traditional recommender systems adopt collaborative filtering techniques to deal with social context. However, they turn out to be computationally intensive and thereby less scalable with internet and social media as input channel. A possible solution is to implement clustering techniques to limit the data to be considered for recommendation process. In tourism environment, based on social media interactions like reviews, forums, blogs, feedbacks, etc. travelers can be clustered to form different interest groups. This experimental study aims at comparing key clustering algorithms with the aim of finding an optimal option that can be adopted in tourism domain by applying social media datasets from travel and tourism context.