您的位置:山东大学 -> 科技期刊社 -> 《山东大学学报(理学版)》

《山东大学学报(理学版)》 ›› 2020, Vol. 55 ›› Issue (11): 35-45.doi: 10.6040/j.issn.1671-9352.1.2019.017

•   • 上一篇    下一篇

基于异质网络嵌入的学术论文推荐方法

许侃(),刘瑞鑫,林鸿飞,刘海峰,冯娇娇,李家平,林原*(),徐博   

  1. 大连理工大学计算机科学与技术学院,辽宁 大连 116024
  • 收稿日期:2019-10-14 出版日期:2020-11-20 发布日期:2020-11-17
  • 通讯作者: 林原 E-mail:xukan@dlut.edu.cn;zhlin@dlut.edu.cn
  • 作者简介:许侃(1981—),男,博士,高级工程师,研究方向为信息检索. E-mail:xukan@dlut.edu.cn
  • 基金资助:
    国家自然科学基金资助项目(61976036);国家自然科学基金资助项目(61602078);教育部人文社会科学研究基金资助项目(16YJCZH128);中央高校基本科研业务费资助项目(DUT19JC45)

Academic paper recommendation based on heterogeneous network embedding

Kan XU(),Rui-xin LIU,Hong-fei LIN,Hai-feng LIU,Jiao-jiao FENG,Jia-ping LI,Yuan LIN*(),Bo XU   

  1. School of Computer Science and Technology, Dalian University of Technology, Dalian 116024, Liaoning, China
  • Received:2019-10-14 Online:2020-11-20 Published:2020-11-17
  • Contact: Yuan LIN E-mail:xukan@dlut.edu.cn;zhlin@dlut.edu.cn

摘要:

针对基于异质信息网络推荐中的有效信息提取与利用,提出了一种基于异质网络嵌入的学术论文推荐方法。使用由元路径引导的随机游走策略生成节点序列;对于每个元路径,通过最大化序列中相邻节点的共现概率来学习节点的唯一嵌入表示;设计了不同的融合函数,将节点在多个不同元路径的低维表示融合为异质信息网络的嵌入,并且引入注意力机制应用于推荐系统。该方法解决了大多数基于异质信息网络的推荐方法因依赖于基于路径的相似性而无法完全挖掘用户和项目潜在结构特征的问题,在DBLP数据集中验证了模型的有效性,并在RMSE指标中取得超过传统模型的效果。

关键词: 异质信息网络, 元路径, 网络嵌入, 论文推荐

Abstract:

This paper focuses on the extraction and utilization of information from heterogeneous information network recommendation, and proposes a heterogeneous network embedding-based academic paper recommendation method. The proposed method uses the random walk strategy guided by the meta-path to generate the sequence of nodes. Then, for each meta-path, the unique embedding representation of nodes is learned by maximizing its co-occurrence probability with the adjacent nodes in the sequence obtained from the given meta-path. In addition, in order to fuse the low-dimensional representation of nodes in differentmetapaths into the final output of heterogeneous information network embedding, the different fusion functions are designed to achieve this goal, and attention mechanism is introduced to recommendation system. This paper applies the heterogeneous information network based on meta paths to solve the problem that most recommendation methods based on heterogeneous information networks rely on path similarity, which can not fully mine the potential structural characteristics of users and projects. We use DBLP data set to validate the effectiveness of the proposed model, which achieved better performance compared with the traditional model on RMSE.

Key words: heterogeneous information network, meta path, network embedding, paper recommend

中图分类号: 

  • TP391.1

图1

本文模型基本原理"

图2

注意力模型的基本原理"

图3

学术论文信息网络中基于元路径的随机游走策略的基本原理"

表1

DBLP数据集"

实体或关系名称 数量
作者 14 475
论文 14 376
作者标签 4
会议 20
论文类型 8 920
作者-标签 4 057
论文-作者 41 794
论文-会议 14 376
论文-类型 114 624

图4

迭代步长对收敛的影响"

图5

参数α、β对实验结果的影响"

图6

潜在因子的维度d对实验结果的影响"

图7

学习率与融合函数对实验结果的影响"

表2

不同模型的比较结果"

学习率/% PMF libFM AttHercs AttHercal AttHercasl
80 0.502 8 0.421 9 0.560 5 0.446 4 0.420 6
60 0.662 6 0.472 6 0.631 0 0.509 6 0.462 4
40 0.800 2 0.556 1 0.704 0 0.554 8 0.566 0
20 0.988 6 0.673 2 0.825 0 0.712 0 0.698 8
1 刘辉, 郭梦梦, 潘伟强. 个性化推荐系统综述[J]. 常州大学学报(自然科学版), 2017, 29 (3): 51- 59.
LIU Hui , GUO Mengmeng , PAN Weiqiang . Overview of personalized recommendation systems[J]. Journal of Changzhou University(Natural Science Edition), 2017, 29 (3): 51- 59.
2 SU Xiaoyuan, KHOSHGOFTAAR T M. A survey of collaborative filtering techniques[J]. Advances in Artificial Intelligence, 2009, http://dx.doi.org/10.1155/2009/421425.
3 王雪蓓.基于协同过滤的个性化推荐[D].广州:华南理工大学, 2016.
WANG Xuebei. The personalized recommendation based on collaborative filtering[D]. Guangzhou: South China University of Technology, 2016.
4 YU Xiao, REN Xiang, GU Quanquan, et al. Collaborative filtering with entity similarity regularization in heterogeneous information networks[C/OL]. 2013, http://hanj.cs.illinois.edu/pdf/hina13_xyu.pdf.
5 YU Xiao, SUN Yizhou, NORICK B, et al. User guided entity similarity search using meta-path selection in heterogeneous information networks[C]//Proceedings of the 21st ACM International Conference on Information and Knowledge Management-CIKM2012, Maui: ACM, 2012: 2025-2029.
6 SUN Yizhou , HAN Jiawei , YAN Xifeng , et al. PathSim: meta path-based top-K similarity search in heterogeneous information networks[J]. Proceedings of the VLDB Endowment, 2011, 4 (11): 992- 1003.
doi: 10.14778/3402707.3402736
7 YU Xiao, REN Xiang, SUN Yizhou, et al. Personalized entity recommendation: a heterogeneous information network approach[C]//Proceedings of the 7th ACM International Conference on Web Search and Data Mining. New York: ACM, 2014: 283-292.
8 SHI Chuan, ZHANG Zhiqiang, LUO Ping, et al. Semantic path based personalized recommendation on weighted heterogeneous information networks[C]//Proceedings of the 24th ACM International on Conference on Information and Knowledge Management. Melbourne: ACM, 2015: 453-462.
9 SHI Chuan , HU Binbin , ZHAO Waynexin , 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
10 GOYAL P , FERRARA E . Graph embedding techniques, applications, and performance: a survey[J]. Knowledge-Based Systems, 2018, 151 (1): 78- 94.
11 PEROZZI B, AL-RFOU R, SKIENA S. Deepwalk: online learning of social representations[C]//Proceedings of 20th International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2014: 701-710.
12 GROVER A, LESKOVEC J. Node2vec: scalable feature learning for networks[C]//Proceedings of the 22nd International Conference on Knowledge Discovery and Data Mining. San Francisco: ACM, 2016: 855-864.
13 SALAKHUTDINOV R, MNIH A. Probabilistic matrix factorization[C]//International Conference on Neural Information Processing Systems. Vancouver: ACM, 2007: 1257-1264.
14 RENDLE S . Factorization machines with libFM[J]. ACM Transactions on Intelligent Systems and Technology, 2012, 3 (3): 1- 22.
[1] 谭红叶,要一璐,梁颖红. 基于知识脉络的科技论文推荐[J]. 山东大学学报(理学版), 2016, 51(5): 94-101.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 杨军. 金属基纳米材料表征和纳米结构调控[J]. 山东大学学报(理学版), 2013, 48(1): 1 -22 .
[2] 何海伦, 陈秀兰*. 变性剂和缓冲系统对适冷蛋白酶MCP-01和中温蛋白酶BP-01构象影响的圆二色光谱分析何海伦, 陈秀兰*[J]. 山东大学学报(理学版), 2013, 48(1): 23 -29 .
[3] 赵君1,赵晶2,樊廷俊1*,袁文鹏1,3,张铮1,丛日山1. 水溶性海星皂苷的分离纯化及其抗肿瘤活性研究[J]. J4, 2013, 48(1): 30 -35 .
[4] 孙小婷1,靳岚2*. DOSY在寡糖混合物分析中的应用[J]. J4, 2013, 48(1): 43 -45 .
[5] 罗斯特,卢丽倩,崔若飞,周伟伟,李增勇*. Monte-Carlo仿真酒精特征波长光子在皮肤中的传输规律及光纤探头设计[J]. J4, 2013, 48(1): 46 -50 .
[6] 杨伦,徐正刚,王慧*,陈其美,陈伟,胡艳霞,石元,祝洪磊,曾勇庆*. RNA干扰沉默PID1基因在C2C12细胞中表达的研究[J]. J4, 2013, 48(1): 36 -42 .
[7] 冒爱琴1, 2, 杨明君2, 3, 俞海云2, 张品1, 潘仁明1*. 五氟乙烷灭火剂高温热解机理研究[J]. J4, 2013, 48(1): 51 -55 .
[8] 杨莹,江龙*,索新丽. 容度空间上保费泛函的Choquet积分表示及相关性质[J]. J4, 2013, 48(1): 78 -82 .
[9] 李永明1, 丁立旺2. PA误差下半参数回归模型估计的r-阶矩相合[J]. J4, 2013, 48(1): 83 -88 .
[10] 董伟伟. 一种具有独立子系统的决策单元DEA排序新方法[J]. J4, 2013, 48(1): 89 -92 .