Existing social networking services suggest friends to customers primarily based on their social graphs, which won’t be the most appropriate to refelect a customer’s preferences on friend selection in real lifestyles. On this project, we present Friendbook, a novel semantic-based friend recommendation system for social networks, which recommends friends to customers based totally on their life styles instead of their social graphs. By taking advantage of sensor-rich smartphones, Friendbook discovers existence types of users from consumer-centric sensor data, measures the similarity of life styles between customers, and recommends friends to customers if their lifestyles patterns have high similarity. Inspired by means of text mining, we model a consumer’s daily life as lifestyles files, from which his/her life patterns are extracted by means of using the Latent Dirichlet Allocation algorithm. We in addition suggest a similarity metric to degree the similarity of lifestyles patterns between users, and calculate customers’ impact in terms of existence styles with a friend-matching graph. Upon receiving a request, Friendbook returns a listing of peoples with maximum recommendation scores to the question consumer. subsequently, Friendbook integrates a feedback mechanism to in addition improve the recommendation accuracy. we’ve got carried out Friendbook on the Android-primarily based smartphones, and evaluated its overall performance on each small-scale experiments and huge-scale simulations. The results show that the recommendation accurately reflect the preferences of users in choosing friends.
DOWNLOAD: NET Project On Friendbook