In this study, we investigated feature-based approaches for improving the link prediction performance for location-based social networks (LBSNs) and analysed their performances. We developed new features based on time, common friend detail and place category information of check-in data in order to make use of information in the data which cannot be utilised by the existing features from the literature. We proposed a feature selection method to determine a feature subset that enhances the prediction performance with the removal of redundant features by clustering them. After clustering features, a genetic algorithm is used to determine the ones to select from each cluster. A non-monotonic and feasible feature selection is ensured by the proposed genetic algorithm. Results depict that both new features and the proposed feature selection method improved link prediction performance for LBSNs.
RoickOHeuserS. Location based social networks: definition, current state of the art and research agenda. Trans GIS2013; 17(5): 763–784.
2.
AggarwalCC. An introduction to social network data analytics. Boston, MA: Springer, 2011, pp. 1–15.
3.
Liben-NowellDKleinbergJ. The link prediction problem for social networks. J Am Soc Inf Sci Technol2007; 58(7): 1019–1031.
4.
ShibataNKajikawaYSakataI. Link prediction in citation networks. J Am Soc Inf Sci Technol2012; 63(1): 78–85.
5.
MatsuoYYamamotoH. Community gravity: measuring bidirectional effects by trust and rating on online social networks. In: Proceedings of the 18th international conference on World Wide Web, Madrid, 20–24 April 2009, pp. 751–760. New York: ACM.
6.
AllamanisMScellatoSMascoloC. Evolution of a location-based online social network: analysis and models. In: Proceedings of the conference on internet measurement conference, Boston, MA, 14–16 November 2012, pp. 145–158. New York: ACM.
7.
SaltonGMcGillMJ. Introduction to modern information retrieval. New York, McGraw-Hill Education, 1983.
8.
NewmanME. Clustering and preferential attachment in growing networks. Phys Rev E2001; 64(2): 025102.
9.
AdamicLAAdarE. Friends and neighbors on the web. Soc Netw2003; 25(3): 211–230.
10.
KatzL. A new status index derived from sociometric analysis. Psychometrika1953; 18(1): 39–43.
11.
ZhengYZhangLMaZ, et al. Recommending friends and locations based on individual location history. ACM T Web2011; 5(1): 5.
12.
ScellatoSNoulasAMascoloC. Exploiting place features in link prediction on location-based social networks. In: Proceedings of the 17th international conference on knowledge discovery and data mining, San Diego, CA, 21–24 August 2001, pp. 1046–1054. New York: ACM.
13.
ChengZCaverleeJLeeK, et al. Exploring millions of footprints in location sharing services. In: Proceedings of the AAAI Conference on Social Media and Weblogs (ICWSM ‘11), Barcelona, Spain, 17–21 July 2011; 81–88. Palo Alto, CA: AAAI.
MengshoelOJDesaiRChenA, et al. Will we connect again? Machine learning for link prediction in mobile social networks. In: Proceedings of the 11th workshop on mining and learning with graphs, Chicago, IL, 11–14 August 2013.
16.
DashMLiuH. Feature selection for classification. Intell Data Anal1997; 1(3): 131–156.
17.
YuLLiuH. Efficient feature selection via analysis of relevance and redundancy. J Mach Learn Res2004; 5: 1205–1224.
18.
BattitiR. Using mutual information for selecting features in supervised neural net learning. IEEE T Neur Netw1994; 5(4): 537–550.
19.
PengHLongFDingC. Feature selection based on mutual information criteria of max-dependency, max-relevance, and min-redundancy. IEEE T Pattern Anal Mach Intell2005; 27(8): 1226–1238.
20.
PrakashMMurtyMN. Feature selection to improve classification accuracy using a genetic algorithm. J Ind Inst Sci2013; 77(1): 85.
21.
XuZJinRYeJ, et al. Non-monotonic feature selection. In: Proceedings of the 26th annual international conference on machine learning, Montreal, Quebec, 14–18 June 2009, pp. 1145–1152. New York: ACM.
22.
ButterworthRPiatetsky-ShapiroGSimoviciDA. On feature selection through clustering. In: Proceedings of the 13th international conference on data mining, Houston, TX, 27–30 November 2005, pp. 581–584. New York: IEEE Computer Society.
23.
SongQNiJWangG. A fast clustering-based feature subset selection algorithm for high-dimensional data. IEEE T Knowl Data Eng2013; 25(1): 1–14.
24.
BayrakAEPolatF. Contextual feature analysis to improve link prediction for location based social networks. In: Proceedings of the 8th workshop on social network mining and analysis, New York, 24–27 August 2014, pp. 1–4. New York: ACM.
25.
BreimanL. Random forests. Mach Learn2001; 45(1): 5–32.
26.
FawcettT. ROC graphs: notes and practical considerations for researchers. Mach Learn2004; 31: 1–38.
27.
LichtenwalterRNLussierJTChawlaNV. New perspectives and methods in link prediction. In: Proceedings of the 16th international conference on knowledge discovery and data mining, Washington, D.C., 25–28 July 2010, pp. 243–252. New York: ACM.
28.
HallMFrankEHolmesG, et al. The WEKA data mining software: an update. ACM SIGKDD Explor Newsl2009; 11(1): 10–18.
