Travel Route Recommendation System

Abstract

The geotagged photos enable people to share their personal experiences as tourists at specific locations and time. Assuming that the collection of each user‟s geotagged photos is a sequence of visited locations, photo-sharing sites are important sources to collect the footprint of tourists. By carefully analysing their footprint movements, our objective is to extract representative and diverse travel routes.

In this paper, we propose a travel route recommendation method that makes use of (1) the travel experiences extracted from Flickr geotagged photos and (2) the landmark from Open Street Map. 639,680 geotagged photos are included in Flickr and 9,227 locations are included in Open Street Map.

We first propose a mapping algorithm to map a geotagged photo is to a landmark based on their longitude and latitude. This mapping enables us to extract important landmarks from our photo collection. Then we use these data to estimate the relationship of the landmarks based on association rule. Lastly, we apply a genetic algorithm to generate tourist routes based on different criterion to satisfy user needs.