Survey Paper on Wayfarer: A Travel Recommendation System

Abstract

In tourism recommendation system, a number of users and items are very large. But traditional recommendation system uses partial information for identifying similar characteristics of users. Collaborative filtering is a primary approach of any recommendation system. We are using the SentiWordNet algorithm for giving ratings to search destinations. Social networking sites act as an interactive platform for today's generation. So a lot can be extracted from social networking sites for recommendations. In the considered system, semantic positive and negative values are being calculated from SentiWordNet. Logged user can apply different filters to search destination. Captions will be extracted through social networking site and upon that text mining will be performed and trending spots will be provided to the customer through a web portal. Real-time data from Instagram is being accessed and process through spring boot. We have created APIs using RESTful web services for a web-based application.