Sustainability-oriented Recommender Systems

Abstract

1 CONTEXT AND MOTIVATION Recommender Systems (RSs) are software tools that provide users with personalised access to products and services, which are collectively referred to as items [15]. To accomplish this, RSs employ various machine learning techniques to evaluate the relevance score of each item-user pair and then recommend the most relevant items to the corresponding user. In tourism applications, which is my research focus, RSs suggest to users relevant points of interest (POIs), based on the user’s profile, context, and preferences [14]. Despite its benefits, this user-focused recommendation approach in tourism domain (where RS acts only on behalf of the users) may have long-term negative impacts on the environment and local communities. This is because this approach encourages the development of user-driven tourism, which is known to have detrimental effects on natural and non-natural landscapes and wildlife habitats where tourism takes place [11]. Furthermore, such an approach may even affect tourists’ and local residents’ experiences; for example, with increased noise levels, traffic congestion and reduced social distances due to overcrowding in tourist areas. Sustainable tourism, like guidelines and management practices, can provide a solution to these issues. Deliberately, it aims to reduce the tensions created by the complex interactions between tourists and the environment in order to mitigate the negative environmental impacts of uncontrolled development in pursuit of economic benefits [3]. Although this concept is well known in the tourism literature, it is relatively new in the RS community. My PhD research involves the integration of sustainability-driven goals into RS design to improve tourism development by providing sustainable recommendations