Recommendation of Attractions and Activities - Using Collaborative Filtering and Implicit Feedback

Abstract

Recommendation systems are becoming more and more popular and are introduced to new domains all of the time. Therefore, the purpose of this master's thesis is to investigate if a recommender system into the domains of attractions and activities is plausible both to create and if it is viable towards the end user. We aim to proof-of-concept a recommendation system which utilizes collaborative filtering, implicit feedback, and user profiling. Further, an experiment is conducted with external users. The experiment will collect data from the users, their thoughts of the service and if they found the service viable and the recommendation given to be credible.

In the first part of the thesis, there is an empirical study of the methods and techniques used in recommendation systems. This study includes methods used in modern recommendation services nowadays and is used as a guideline when creating and discussing the service. The second part of the thesis undertake the requirement, design and implementation parts where we decide which methods to utilize, how we created the service and what technologies used. Lastly, we discuss and evaluate our findings regarding the service's viability and credibility using the results obtained in the conducted experiment.

The results of the proof-of-concept experiments show that the participants found the system credible, and the thesis will argue that such a recommendation system is viable.