×

Effects of the high-order correlation on information filtering. (English) Zbl 06199302

Summary: In this paper, we empirically investigate the statistical properties of the user correlation network in terms of their common rated objects on MovieLens, and find that it has high clustering coefficient and ultra small average distance, which is close to the fully connected network. We argue that the above characteristics come from the fact that large-degree objects build lots of fully connected subnetworks by using the node projection method. By introducing the user global similarity, measured by the product of two users’ similarity vectors, we present an effective way to identify users’ specific interests by weakening the mainstream interests and noise interests. Numerical results show that we are able to obtain accurate and diverse recommendations by considering the second-order correlation redundant information simultaneously, which outperforms the state-of-the-art collaborative filtering (CF) methods. This work suggests that statistical properties of the user correlation network is an important factor to improve the performances of information filtering algorithms.

MSC:

62-XX Statistics
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] DOI: 10.1038/30918 · Zbl 1368.05139 · doi:10.1038/30918
[2] Barabási A.-L., Science 286 pp 509–
[3] DOI: 10.1103/RevModPhys.74.47 · Zbl 1205.82086 · doi:10.1103/RevModPhys.74.47
[4] DOI: 10.1080/00018730110112519 · doi:10.1080/00018730110112519
[5] DOI: 10.1137/S003614450342480 · Zbl 1029.68010 · doi:10.1137/S003614450342480
[6] DOI: 10.1142/S0218127402004802 · Zbl 1044.37561 · doi:10.1142/S0218127402004802
[7] DOI: 10.1016/j.physrep.2005.10.009 · Zbl 1371.82002 · doi:10.1016/j.physrep.2005.10.009
[8] DOI: 10.1016/j.physa.2005.10.036 · doi:10.1016/j.physa.2005.10.036
[9] DOI: 10.1103/PhysRevE.66.065102 · doi:10.1103/PhysRevE.66.065102
[10] DOI: 10.1103/PhysRevLett.91.148701 · doi:10.1103/PhysRevLett.91.148701
[11] DOI: 10.1103/PhysRevE.74.056109 · doi:10.1103/PhysRevE.74.056109
[12] DOI: 10.1103/PhysRevE.72.026133 · doi:10.1103/PhysRevE.72.026133
[13] DOI: 10.1142/S0217984905008773 · Zbl 1078.90017 · doi:10.1142/S0217984905008773
[14] Jin F., Phys. A 355 pp 657–
[15] DOI: 10.1142/S0217984906010883 · Zbl 1103.68342 · doi:10.1142/S0217984906010883
[16] DOI: 10.1103/PhysRevE.76.046115 · doi:10.1103/PhysRevE.76.046115
[17] DOI: 10.1145/963770.963775 · Zbl 05457973 · doi:10.1145/963770.963775
[18] DOI: 10.1145/316194.316229 · doi:10.1145/316194.316229
[19] DOI: 10.1016/S1389-1286(00)00083-9 · doi:10.1016/S1389-1286(00)00083-9
[20] DOI: 10.1109/TKDE.2005.99 · Zbl 05109798 · doi:10.1109/TKDE.2005.99
[21] DOI: 10.1145/963770.963772 · Zbl 05457966 · doi:10.1145/963770.963772
[22] DOI: 10.1145/245108.245126 · doi:10.1145/245108.245126
[23] DOI: 10.1103/PhysRevE.85.016118 · doi:10.1103/PhysRevE.85.016118
[24] DOI: 10.1073/pnas.1000488107 · doi:10.1073/pnas.1000488107
[25] DOI: 10.1016/j.physa.2009.10.027 · doi:10.1016/j.physa.2009.10.027
[26] DOI: 10.1103/PhysRevE.80.017101 · doi:10.1103/PhysRevE.80.017101
[27] DOI: 10.1142/S0129183109013613 · Zbl 1184.68493 · doi:10.1142/S0129183109013613
[28] DOI: 10.1103/PhysRevLett.99.154301 · doi:10.1103/PhysRevLett.99.154301
[29] DOI: 10.1287/mnsc.1060.0619 · Zbl 1232.91426 · doi:10.1287/mnsc.1060.0619
[30] DOI: 10.1209/0295-5075/90/48006 · doi:10.1209/0295-5075/90/48006
[31] DOI: 10.1103/PhysRevE.72.036120 · doi:10.1103/PhysRevE.72.036120
[32] DOI: 10.1103/PhysRevE.71.026108 · doi:10.1103/PhysRevE.71.026108
[33] DOI: 10.1016/j.physa.2006.04.047 · doi:10.1016/j.physa.2006.04.047
[34] DOI: 10.1016/j.dam.2008.06.052 · Zbl 1182.92032 · doi:10.1016/j.dam.2008.06.052
[35] DOI: 10.1016/j.socnet.2007.04.006 · doi:10.1016/j.socnet.2007.04.006
[36] DOI: 10.1103/PhysRevE.76.066102 · doi:10.1103/PhysRevE.76.066102
[37] DOI: 10.1103/PhysRevLett.89.208701 · doi:10.1103/PhysRevLett.89.208701
[38] DOI: 10.1103/PhysRevLett.96.114102 · doi:10.1103/PhysRevLett.96.114102
[39] DOI: 10.1073/pnas.98.2.404 · doi:10.1073/pnas.98.2.404
[40] DOI: 10.1103/PhysRevE.64.016131 · doi:10.1103/PhysRevE.64.016131
[41] DOI: 10.1016/S0378-4371(02)00736-7 · Zbl 0996.91086 · doi:10.1016/S0378-4371(02)00736-7
[42] DOI: 10.1017/CBO9780511610905 · doi:10.1017/CBO9780511610905
[43] DOI: 10.1103/PhysRevE.73.036132 · doi:10.1103/PhysRevE.73.036132
[44] Wasserman S., Social Networks Analysis (2001)
[45] Scott J., Social Network Analysis: A Handbook (2000)
[46] DOI: 10.1103/PhysRevE.72.056127 · doi:10.1103/PhysRevE.72.056127
[47] DOI: 10.1088/1367-2630/9/7/228 · doi:10.1088/1367-2630/9/7/228
[48] DOI: 10.1209/0295-5075/89/18001 · doi:10.1209/0295-5075/89/18001
[49] DOI: 10.1103/PhysRevE.84.037101 · doi:10.1103/PhysRevE.84.037101
[50] DOI: 10.1023/A:1009804230409 · Zbl 1006.68628 · doi:10.1023/A:1009804230409
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.