iGSLR: personalized geo-social location recommendation: a kernel density estimation approach
Citations
582 citations
520 citations
Cites background from "iGSLR: personalized geo-social loca..."
...In terms of improving the efficiency of the location recommendations [20, 107], Chow et al., propose a new recommendation algorithm that using the safe region technique to reduce the system communicational and computational overhead for the users moving on their paths....
[...]
...Zhang and Chow [125] further explore the geographical influences in location recommendation, from the perspective of a user’s personalized travel pattern....
[...]
...These systems can be divided into two groups by the objectives of their recommendation: 1) stand-alone location recommender systems, which Table 3 Summary of the existing recommender systems in location-based social networks Objectives Methodologies Data sources Social Location User Activity Content Link CF User Individual User Media based analysis profile locations trajectories Sandholm [83] √ √ √ Levandoski [55] √ √ √ √ Park [76] √ √ √ Horozov [45] √ √ √ Ye [111] √ √ √ Chow [20] √ √ √ Ye [112] √ √ √ Tai [94] √ √ √ Yoon [121] √ √ √ Cao [13] √ √ √ Ye [110] √ √ √ Liu [62] √ √ √ Zheng [136] √ √ √ √ Zheng [133] √ √ √ √ Li [56] √ √ √ Hung [47] √ √ √ Xiao [106] √ √ √ √ Ying [120] √ √ √ √ Scellato [87] √ √ √ Zheng [127] √ √ √ √ √ Symeonidis [75] √ √ √ √ Yin [114] √ √ √ √ provide a user with individual locations, such as restaurants or cities, that match their preferences, and 2) sequential location recommender systems, which recommend a series of locations (e.g., a popular travel route in a city) to a user based on their preferences and their constraints, such as in time and cost....
[...]
310 citations
Cites methods from "iGSLR: personalized geo-social loca..."
...[32] developed a unified geo-social recommendation framework, namely iGSLR, in which a kernel density estimation approach was used to personalize the geographical influence on users’ check-in behaviors as individual distributions....
[...]
299 citations
290 citations
References
15,499 citations
11,727 citations
"iGSLR: personalized geo-social loca..." refers methods in this paper
...Noth that: (1) in iGSLR the the geographical influence of locations is personalized, as presented in Section 3; (2) the model parameters of PD are obtained using maximum likelihood estimation; and (3) the centers of MGM are discovered by the mean-shift clustering algorithm [4]....
[...]
8,634 citations
"iGSLR: personalized geo-social loca..." refers methods in this paper
...Memory-based methods can be further grouped into user-based CF [10] and item-based CF [19]....
[...]
...or (ii) the item-based CF method [19] (“item” means “location” in the case of LBSNs):...
[...]
2,922 citations
"iGSLR: personalized geo-social loca..." refers background or methods in this paper
...Finally, we have conducted extensive experiments to evaluate the performance of iGSLR using two large-scale real data sets collected from Foursquare and Gowalla....
[...]
...Recently with the emergence of LBSNs, like Foursquare, Gowalla, and Facebook places, recommending locations (i.e., POIs) for users becomes prevalent....
[...]
...(Section 4) • We conduct extensive experiments to evaluate the performance of iGSLR using two large-scale real data sets collected from Foursquare and Gowalla....
[...]
...However, the improvement of SCF could be considerably limited, because in general users with social friendships only share less than 10% commonly visited locations [2, 3, 24]....
[...]
...(2) On the Gowalla data set with one-order-of-magnitude lower density, iGSLR still outperforms the other recommendation methods to a large extent, whereas PD deteriorates dramatically, even worse than SG and MGM. (3) Since it is important for LBSNs to provide good recommendation for cold-start users, iGSLR is better than other state-of-the-art geo-social recommendation techniques for LBSNs to attract new users....
[...]
2,686 citations