Exploiting POI-Specific Geographical Influence for Point-of-Interest Recommendation

Abstract

Point-of-interest (POI) recommendation, i.e., recommending unvisited POIs for users, is a fundamental problem for location-based social networks. POI recommendation distinguishes itself from traditional item recommendation, e.g., movie recommendation, via geographical influence among POIs. Existing methods model the geographical influence between two POIs as the probability or propensity that the two POIs are co-visited by the same user given their physical distance. These methods assume that geographical influence between POIs is determined by their physical distance, failing to capture the asymmetry of geographical influence and the high variation of geographical influence across POIs. In this paper, we exploit POI-specific geographical influence to improve POI recommendation. We model the geographical influence between two POIs using three factors: the geo-influence of POI, the geo-susceptibility of POI, and their physical distance. Geo-influence captures POI?s capacity at exerting geographical influence to other POIs, and geo-susceptibility reflects POI?s propensity of being geographically influenced by other POIs. Experimental results on two real-world datasets demonstrate that POI-specific geographical influence significantly improves the performance of POI recommendation.

Modeling Users’ Multifaceted Interest Correlation for Social Recommendation

“…We use two real-world datasets collected from Gowalla [20] and Epinions [12] respectively for evaluation.…”

“…|x| denotes the cardinality of set x. For each metric, we consider 4 values (i.e., 1,5,10,20) of n in our experiments.…”

Topic-Enhanced Memory Networks for Personalised Point-of-Interest Recommendation

“…Vanilla MF was initially used to deal with general POI recommendation, where only user-POI interactions are leveraged. When additional information like spatio-temporal information and social relationship became available, some models derived from basic MF (e.g., IRenMF [19], GeoMF [17], GeoIE [25]) were proposed and adapted to context-aware recommendation. Markov chain models (e.g., LBPR [11], NLPMM [4], FPMC-LR [6], LORE [32]) and recurrent neural network (RNN) (e.g., ST-RNN [18], CARA [20], DRCF [18]) have also been employed when temporal effects are particularly valued in sequential POI recommendation.…”

“…Unlike other recommender systems that push digital goods, e.g. e-books, news, and movies, POI recommendation aims at offering users preferred new venues to explore in the physical world [25], which can be largely affected by various real-life factors and thus faces more challenges. Firstly, to experience a POI, a user has to physically visit it, which is generally more costly and timeconsuming than watching a movie or listening to a song online.…”

“…For these reasons, the number of POIs that a user generally interacts with is extremely small compared with the total number of POIs in the LBSN, leading to a very sparse user-POI matrix that plagues a considerable number of current POI recommender systems [29]. Moreover, unlike pure online interactions, users' activities in the physical world are limited by travel distance and time [25], making POI recommendation a more complex issue. Considering the distinguishing characteristics of POI recommendation task, corresponding research efforts have been spend improving its effectiveness.…”