FindToME :Un Système de Recommandation Collaboratif de Lieux

Abstract

This project is positioned in the field of recommender systems (RS). A recommendation systemis an information search and filtering tool that aims to pro- pose to users items that propose to users items that may be of interest. Globally, RS are based on the analysis of users preferences and their implicit or explicit evaluations for items. Searching for and finding a good doctor, restaurant, school, dentist, cafeteria, etc. in real time from one’s current from one’s current location remains difficult to achieve in Algeria. In order to facilitate to the Algerian user access to different places. We have developed a collaborative recommendation system called FindToMe which learns to classify the different places according to their distances and the opinions of users reviews. A step of collecting users opinions and feedback has been conducted on the different places in the city of Guelma. Then, we applied an analysis of the data collected based on the VADER sentiment analysis method. Then, for the recommendation of the interesting places a recommendation algorithm is designed and implemented according to the following four (04) parameters : the location, the rating, the comments and the age. The results we obtained are encou- raging and allow us to conclude that that our recommendation system succeeds in suggesting items that correspond to the users preferences