Point-of-Interest Recommendation Algorithm Combining Popularity Feature and Kernel Density Estimation

Ya-li SI,Feng LI,Ya-wei SONG
2016-01-01
Abstract:Point-of-interest (POI)recommendation in Location-Based Social Networks (LBSNs) can not only help users to explore new locations and bring experiences effectively,but.also help shoppers to send advertisements and improve business benefits.Aiming at the low accuracy of POI recommendation,this paper proposes a point-of-interest recommendation algorithm combined with popularity feature and kernel density estimation.Firstly,the popularity feature of POI and time continuity feature are extracted seperately from the large-scale check-in dataset in LBSNs.Then,the POI popularity evaluation method based on continuous time slot is proposed.On this basis,we design a POI recommendation algorithm called PKDE,which utilizing two-dimensional Gauss kernel density estimation,and finally we evaluate the algorithm through experimental comparison and analysis based on three large-scale real LBSNs data sets.Experimental results show that the proposed algorithm can effectively improve the recommendation precision and recall compared with the existing similar algorithms.
What problem does this paper attempt to address?