Abstract
With the rapid development of the Internet, high-dimensional, sparse, and redundant data often appear in complex networks. These data require effective link prediction techniques to extract the most basic and relevant information for online user services. In this paper, we propose a link prediction algorithm based on a direct optimization of the AUC (area under the curve). In the proposed algorithm, the AUC is treated as the objective function for optimization, and link prediction is transformed into a binary classification problem, where the class label of each node pair is determined by whether there exists a direct link between them. The binary classification problem can then be solved by AUC optimization. We use the hinge function as the loss function and iteratively update the weight matrix based on the stochastic gradient sub-descent method. We test our method on several real-world heterogeneous information networks that are chosen from different domains and are diverse in structure and relationship type. The empirical results show that our algorithm can achieve higher quality prediction results than those of other algorithms.
Similar content being viewed by others
References
Lichtenwalter RN (2010) New precepts and method in link prediction. In: Proceedings of ACM KDD’10, pp 243–252
Lü LY, Zhou T (2011) Link prediction in complex networks: a survey. Physica A 390:1150–1170
Kaya B, Poyraz M (2015) Age-series based link prediction in evolving disease networks. Comput Biol Med 63:1–10
Zhang L, Hu K, Tang Y (2010) Predicting disease-related genes by topological similarity in human protein-protein interaction network. Cent Eur J Phys 8(4):672–682
Guimera R, Sales-Pardo M (2010) Missing and spurious interactions and the reconstruction of complex networks. Proc Natl Acad Sci USA 106(52):22073–22078
Papadimitriou A, Symeonidis P, Manolopoulos Y (2012) Fast and accurate link prediction in social networking systems. J Syst Softw 85(9):2119–2132
Buccafurri F, Lax G, Nocera A, Ursino D (2015) Discovering missing me edges across social networks. Inf Sci 319:18–37
Fournet J, Barrat A (2014) Contact patterns among high school students. PLoS ONE 9(9):e107878
Hossmann T, Nomikos G, Spyropoulos T, Legendre F (2012) Collection and analysis of multi-dimensional network data for opportunistic networking research. Comput Commun 35(13):1613–1625
Ibrahim NMA, Chen L (2015) Link prediction in dynamic social networks by integrating different types of information. Appl Intell 42(4):738–750
Jahanbakhsh K, King V, Shoja GC (2012) Predicting missing contacts in mobile social networks. Pervasive Mob Comput 8(5):698–716
Sun Y, Barbery R, Gupta M, Aggarwal CC, Han J (2011) Co-author relationship prediction in heterogeneous bibliographic networks. In: Proceedings of 2011 international conference on advances in social networks analysis and mining (ASONAM 2011), pp 121–128
Li J, Zhang LL, Meng F, Li FH (2014) Recommendation algorithm based on link prediction and domain knowledge in retail transactions. Procedia Computer Science 31:875– 881
Xie F, Chen Z, Shang JX, Feng XP, Li J (2015) A link prediction approach for item recommendation with complex number. Knowl-Based Syst 81:148–158
Li X, Chen HC (2013) Recommendation as link prediction in bipartite graphs: a graph kernel-based machine learning approach. Decis Support Syst 54(2):880–890
Vidmer A, Zeng A, Medo M, Zhang YC (2015) Prediction in complex systems: The case of the international trade network. Physica A: Statistical Mechanics and its Applications 436:188– 199
Huang Z, Lin DKJ (2009) The time-series link prediction problem with applications in communication surveillance. INFORMS J Comput 21:286–303
Kaya B, Poyraz M (2014) Supervised link prediction in symptom networks with evolving case. Measurement 56:231–238
Liu HK, Lü LY, Zhou T (2011) Uncovering the network evolution mechanism by link prediction (in Chinese). Sci Sin Phys Mech Astron 41:816–823
Ma X, Liao JL, Djouadi SM, Cao Q (2014) LIPS: link prediction as a service for data aggregation applications. Ad Hoc Netw 19:43–58
Yang Y, Chawla N, Sun Y, Han JW (2012) Predicting links in multi-relational and heterogeneous networks. In: 2012 IEEE 12th international conference on data mining, pp 756–764
Davis D, Lichtenwalter R, Chawla NV (2011) Multi-relational link prediction in heterogeneous information networks. In: Proceedings of 2011 international conference on advances in social networks analysis and mining, pp 281–288
Stroele V, Zimbrao G, Souza JM (2013) Group and link analysis of multi-relational scientific social networks. J Syst Softw 86:1819–1830
Ermis B, Cemgil AT (2014) A Bayesian tensor factorization model via variational inference for link prediction[J]. arXiv:1409.8276
Zhang T (2004) Statistical behavior and consistency of classification methods based on convex risk minimization [J]. Ann Stat 32:56–85
Wang J, Tao Q (2008) Machine learning: the state of the art [J]. IEEE Intell Syst 23:49–55
Yuan GX, Chang KW, Hsieh CJ, Lin CJ (2010) A comparisom of optimization methods and software for large-scale L1-regularized linear classification[J]. J Mach Learn Res 11(2):3183–3234
Sun ZY, Tao Q (2009) A statistical machine learning survey: loss function and optimization [J](in Chinese). Journal of communications of the China Computer Fedaration 5:7–14
Hastie T, Zhu J (2006) Comment [J]. Stat Sci 21:352–357
Kossinets G (2010) Effects of missing data in social networks. Soc Networks 28(3):247–268
Chierichetti F, Kumar R, Pandey S, Vassilvitskii S (2010) Finding the Jaccard median. In: Proceedings of the twenty-first annual ACM-SIAM symposium on discrete algorithms, pp 293–311
Xie YB, Zhou T, Wang BH (2008) Scale-free networks without growth. Physica A 387(7):1683–1688
Adamic LA, Adar E (2003) Friends and neighbors on the web. Soc Networks 25(3):211–230
Lee C, Pham M, Jeong MK, Kim D, Lin DKJ, Chavalitwongse WA (2015) A network structural approach to the link prediction problem. INFORMS J Comput 27:249–267
Wang ZX, Wu Y, Li QG, Jin FD, Xiong W (2016) Link prediction based on hyperbolic mapping with community structure for complex networks. Physica A: Statistical Mechanics and its Applications 450:609–623
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Funding
This research was supported in part by the Chinese National Natural Science Foundation under grant Nos. 61379066, 61070047, 61379064, 61472344, and 61402395; the Natural Science Foundation of Jiangsu Province under contracts BK20130452, BK2012672, BK2012128, and BK20140492; the Natural Science Foundation of the Education Department of Jiangsu Province under contracts 12KJB520019, 13KJB520026, and 09KJB20013;and the Six Talent Peaks Project in Jiangsu Province(Grant No. 2011-DZXX-032).
Conflict of interests
No conflict of interest.
Rights and permissions
About this article
Cite this article
Dai, C., Chen, L. & Li, B. Network link prediction based on direct optimization of area under curve. Appl Intell 46, 427–437 (2017). https://doi.org/10.1007/s10489-016-0845-4
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10489-016-0845-4