A Personalized Paper Recommendation Algorithm Based on Heterogeneous Graph Embedding

doi:10.11871/jfdc.issn.2096-742X.2023.06.14

Abstract

Abstract:

[Background] With the rapid growth of the number of scientific papers, finding or locating the papers of interest has become an urgent problem for researchers in the process of scientific research. [Objective] This paper aims to study a paper recommendation algorithm to solve the problem of user-oriented personalized paper recommendations. [Methods] A personalized paper recommendation algorithm based on heterogeneous graph embedding is proposed, which learns the representation of the paper nodes, then the interest of the author is calculated according to the papers published by the author, and recommendations are then generated based on the similarity between author interests and the papers. [Results] Experiment on the DBLP dataset demonstrates the effectiveness of the model and the algorithm proposed in this paper.

Key words: graph embedding, paper recommendation, heterogeneous graph, personalized recommendation

ZHAO Chengliang, CHEN Yuanping. A Personalized Paper Recommendation Algorithm Based on Heterogeneous Graph Embedding[J]. Frontiers of Data and Computing, 2023, 5(6): 153-160, https://cstr.cn/32002.14.jfdc.CN10-1649/TP.2023.06.014.

Figures/Tables 7

Fig.1

Fig.2

Table 1

Table 2

Table 3

Table 4

Table 5

References 20

[1]	TIAN G, JING L. Recommending Scientific Articles Using Bi-Relational Graph-Based Iterative RWR[C]// Proceedings of the 7th ACM Conference on Recommender Systems, New York, NY, USA:, Association for Computing Machinery, 2013: 399-402.
[2]	GORI M, PUCCI A. Research Paper Recommender Systems: A Random-Walk Based Approach[C]// 2006 IEEE/WIC/ACM International Conference on Web Intelligence (WI 2006 Main Conference Proceedings)(WI'06), 2006: 778-781.
[3]	MENG F, GAO D, LI W, et al. A Unified Graph Model for Personalized Query-Oriented Reference Paper Recommendation[C]// Proceedings of the 22nd ACM International Conference on Information and Knowledge Management, New York, NY, USA:, Association for Computing Machinery, 2013: 1509-1512.
[4]	TOTTI L C, MITRA P, OUZZANI M, et al. A Query-Oriented Approach for Relevance in Citation Networks[C]// Proceedings of the 25th International Conference Companion on World Wide Web. Republic and Canton of Geneva, CHE:, International World Wide Web Conferences Steering Committee, 2016: 401-406.
[5]	PEROZZI B, Al-RFOU R, SKIENA S. DeepWalk: Online Learning of Social Representations[C]// Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, New York, NY, USA:, Association for Computing Machinery, 2014: 701-710.
[6]	SHI C, HU B, ZHAO W X, et al. Heterogeneous Information Network Embedding for Recommendation[J]. IEEE Transactions on Knowledge and Data Engineering, 2019, 31(2): 357-370. doi: 10.1109/TKDE.2018.2833443
[7]	BEEL J, GIPP B, LANGER S, et al. Research-paper recommender systems: a literature survey[J]. International Journal on Digital Libraries, 2016, 17(4): 305-338. doi: 10.1007/s00799-015-0156-0
[8]	YANG C, WEI B, WU J, et al. CARES: A Ranking-Oriented CADAL Recommender System[C]// Proceedings of the 9th ACM/IEEE-CS Joint Conference on Digital Libraries, New York, NY, USA:, Association for Computing Machinery, 2009: 203-212.
[9]	ZHOU Q, CHEN X, CHEN C. Authoritative Scholarly Paper Recommendation Based on Paper Communities[C]// 2014 IEEE 17th International Conference on Computational Science and Engineering, 2014: 1536-1540.
[10]	CAI X, HAN J, LI W, et al. A Three-Layered Mutually Reinforced Model for Personalized Citation Recommendation[J]. IEEE Transactions on Neural Networks and Learning Systems, 2018, 29(12): 6026-6037. doi: 10.1109/TNNLS.2018.2817245 pmid: 29993933
[11]	GUO L, CAI X, HAO F, et al. Exploiting Fine-Grained Co-Authorship for Personalized Citation Recommendation[J]. IEEE Access, 2017, 5: 12714-12725. doi: 10.1109/ACCESS.2017.2721934
[12]	XIA F, LIU H, LEE I, et al. Scientific Article Recommendation: Exploiting Common Author Relations and Historical Preferences[J]. IEEE Transactions on Big Data, 2016, 2(2): 101-112. doi: 10.1109/TBDATA.2016.2555318
[13]	LIU X, YU Y, GUO C, et al. Meta-Path-Based Ranking with Pseudo Relevance Feedback on Heterogeneous Graph for Citation Recommendation[C]// Proceedings of the 23rd ACM International Conference on Conference on Information and Knowledge Management, New York, NY, USA: Association for Computing Machinery, 2014: 121-130.
[14]	REN X, LIU J, YU X, et al. ClusCite: Effective Citation Recommendation by Information Network-Based Clustering[C]// Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, New York, NY, USA: Association for Computing Machinery, 2014: 821-830.
[15]	MA X, ZHANG Y, ZENG J. Newly Published Scientific Papers Recommendation in Heterogeneous Information Networks[J]. Mobile Networks and Applications, 2019, 24(1): 69-79. doi: 10.1007/s11036-018-1133-9
[16]	MA X, WANG R. Personalized Scientific Paper Recommendation Based on Heterogeneous Graph Representation[J]. IEEE Access, 2019, 7: 79887-79894. doi: 10.1109/Access.6287639
[17]	CAI X, HAN J, PAN S, et al. Heterogeneous information network embedding based personalized query-focused astronomy reference paper recommendation[J]. International Journal of Computational Intelligence Systems, 2018, 11(1): 591-599. doi: 10.2991/ijcis.11.1.44
[18]	CAI X, ZHENG Y, YANG L, et al. Bibliographic Network Representation Based Personalized Citation Recommendation[J]. IEEE Access, 2019, 7: 457-467. doi: 10.1109/ACCESS.2018.2885507
[19]	KONG X, MAO M, WANG W, et al. VOPRec: Vector Representation Learning of Papers with Text Information and Structural Identity for Recommendation[J]. IEEE Transactions on Emerging Topics in Computing, 2021, 9(1): 226-237. doi: 10.1109/TETC.6245516
[20]	TANG J, ZHANG J, YAO L, et al. ArnetMiner: Extraction and Mining of Academic Social Networks[C]// KDD'08, 2008: 990-998.

	Train-Small	Train-Big
论文节点	126，648	497，379
作者节点	34，622	196，004
Cite关系	268，516	2，250，742
Write关系	51，063	429，394
Co-work关系	135，236	962，246

参数名称	参数取值
numwalks	20
walklength	100
vecotr_size	128
window	5
sg	1
hs	1
epochs	10

	P@10	P@20	P@30	R@10	R@20	R@30	F1@10	F1@20	F1@30
DeepWalk	0.2213	0.1439	0.1068	0.1267	0.1565	0.1687	0.1546	0.1421	0.1236
node2vec	0.2707	0.1854	0.1407	0.1507	0.1939	0.2133	0.1860	0.1802	0.1608
BFS	0.2950	0.2139	0.1659	0.1646	0.2246	0.2511	0.2030	0.2087	0.1898
BFS-papp	0.3013	0.2184	0.1710	0.1693	0.2296	0.2596	0.2082	0.2133	0.1959
BFS-pp	0.3006	0.2165	0.1687	0.1674	0.2260	0.2546	0.2067	0.2107	0.1928
BFS-app	0.3143	0.2285	0.1766	0.1771	0.2397	0.2687	0.2177	0.2228	0.2026

	P@10	P@20	P@30	R@10	R@20	R@30	F1@10	F1@20	F1@30
DeepWalk	0.2593	0.1833	0.1461	0.1333	0.1807	0.2116	0.1667	0.1703	0.1615
node2vec	0.2941	0.2109	0.1680	0.1501	0.2069	0.2413	0.1883	0.1955	0.1850
BFS	0.3090	0.2276	0.1839	0.1545	0.2177	0.2567	0.1951	0.2082	0.2000
BFS-papp	0.3165	0.2310	0.1854	0.1604	0.2236	0.2625	0.2015	0.2123	0.2025
BFS-pp	0.3177	0.2366	0.1864	0.1643	0.2269	0.2674	0.2046	0.2073	0.2005
BFS-app	0.3384	0.2503	0.2011	0.1716	0.2421	0.2834	0.2155	0.2300	0.2192

方法	Small	Big
DeepWalk	2.39	3.48
node2vec	6.41	18.93
BFS	16.05	16.55
BFS-app	23.85	27.8