Ubiquitous Crowdsourcing Model for Location Recommender System

Crowdsourcing is methodology in which task is completed by distributing it amongst the crowd. The objective of proposed work is to generate recommendation of places close to tourist’s current location using crowdsourcing approach. Technological advances in WWW and the mobile devices have opened the possibilities of gathering and sharing the required information from people moving around. Contextual information from user can be collected by several techniques like sensors, collaborative tagging, crowdsourcing etc. In this work, contextual information about the places is gathered from crowd visiting those places and their collective knowledge is further used to generate recommendations for the tourist. Since crisp quantification of context parameters such as weather, traffic, crowdedness is difficult for a general user (crowd), this information is collected from them in terms of fuzzy linguistic variables and fuzzy inference system is used to generate a popularity score of each place nearby tourist’s current location. Finally, the system sorts the score of each place in order of their popularity score and sends these recommendations to the tourist. Additionally, the proposed system collects latest images, audio recorded feedback etc. from the crowd currently present at to be recommended place. These collected images, audio clips, feedbacks etc may be pushed along with recommendations to help the tourist to take decision about visiting these places. Additional current information about tourist spot definitely improves the quality of recommendation and experience. To implement this concept, a prototype system has been implemented using Android SDK, database is designed using MYSQL, fuzzy inference system is simulated using fuzzy logic toolbox of MATLAB and backend tasks are performed using PHP. The proposed prototype has also been tested over a set of real users.