Evolving recommendations from past travel sequences using soft computing techniques

The World Wide Web (WWW) and mobile devices have become an indispensable part of life in this time. The pervasiveness of location acquisition technologies like global positioning system (GPS) has enabled the convenient logging of the movement sequences of users using mobile devices. This work proposes a personalised tourist spot recommender system for mobile users using genetic algorithm (GA) for a situation when explicit user ratings for tourist spots are not available. Implicit ratings of users for tourist spots are mined using GPS trajectory logs. GA is used to evolve ratings of unvisited spots using implicit ratings. GPS trajectory dataset of 178 users collected by Microsoft Research Asia's GeoLife project is used for the purpose of evaluation and experiments. We emphasise that proposed approach is comparable with existing related approaches when compared in terms of average root mean squared error (RMSE) and provides focused, personalised and relevant recommendations.