[1]潘嘉琪,邹俊韬.一种基于深度 RTRBM 的动态网络链路预测方法[J].计算机技术与发展,2020,30(03):1-6.[doi:10. 3969 / j. issn. 1673-629X. 2020. 03. 001]
　PAN Jia-qi,ZOU Jun-tao.A Dynamic Link Prediction Method Based on Deep Recurrent Temporal Restricted Boltzmann Machine[J].Computer Technology and Development,2020,30(03):1-6.[doi:10. 3969 / j. issn. 1673-629X. 2020. 03. 001]
点击复制

一种基于深度 RTRBM 的动态网络链路预测方法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
30
期数:
2020年03期
页码:
1-6
栏目:
智能、算法、系统工程
出版日期:
2020-03-10

文章信息/Info

Title:
A Dynamic Link Prediction Method Based on Deep Recurrent Temporal Restricted Boltzmann Machine
文章编号:
1673-629X(2020)03-0001-06
作者:
潘嘉琪; 邹俊韬
南京航空航天大学 计算机科学与技术学院,江苏 南京 211106
Author(s):
PAN Jia-qi; ZOU Jun-tao
School of Computer Science and Technology,Nanjing University of Aeronautics and Astronautics, Nanjing 211106,China
关键词:
动态网络; 链路预测; 网络嵌入; 受限玻尔兹曼机
Keywords:
Temporal Network; Link Prediction; Network embedding; Restricted Boltzmann Machine
分类号:
TP311
DOI:
10. 3969 / j. issn. 1673-629X. 2020. 03. 001
摘要:
针对节点对的嵌入特征随时间演化而发生的骤变问题,提出了一种基于深度循环时序受限玻尔兹曼机(RTRBM) 的链路预测方法。 在样本集构建方面,利用网络嵌入学习自动化提取网络节点特征,并以嵌入特征空间中两个节点间的 距离作为节点对样本属性;在学习模型选择方面,将RTRBM模型应用于动态网络链路预测,考虑到短时间间隔内节点在 嵌入特征空间中的位置相对稳定,对RTRBM的能量函数及训练过程进行了改进。 此外,为了提取节点对的深度时序特 征,结合深度学习理论,通过纵向地堆叠多个改进后的RTRBM构成深度学习结构,并利用Logistic回归分类器对动态网络 中的链路关系进行分类和预测。 实验结果表明,改进后的RTRBM及其深度学习模型相比于其他方法在AUC指标下有着 明显的性能提升。
Abstract:
Aiming at the abrupt change of embedded features of node pairs with time evolution,a temporal link prediction method based on deep recurrent temporal restricted Boltzmann machine(RTRBM) is proposed. In the aspect of sample set construction,network embedding is used to extract node features over networks, and the distance between node pairs is taken as sample attributes. In terms of learning model selection,RTRBM is applied to temporal link prediction. Considering the position of the node in the embedded features space is relatively stable in a short time interval, the energy function and training process of the RTRBM are improved. In addition,In order to extract the depth-series features of node pairs,the deep learning theory is combined to form a deep leaning structure by stacking multiple improved RTRBM vertically,and a logistic regression classifier is utilized finally to classify and predict link relationships in dynamic networks. The experimental results show that the improved RTRBM and its deep learning model have significant performance improvement under the AUC index compared with other methods.

相似文献/References:

[1]顾炎,熊超. 动态社区的点增量发现算法[J].计算机技术与发展,2017,27(06):81.
　GU Yan,XIONG Chao. Vertex-based Incremental Algorithm for DynamicCommunities Detecting[J].Computer Technology and Development,2017,27(03):81.
[2]李旗旗,徐敏. 社交网络中的链路预测方法改进[J].计算机技术与发展,2017,27(11):37.
　LI Qi-qi,XU Min. Improvement of Link Prediction Method in Social Networks[J].Computer Technology and Development,2017,27(03):37.
[3]李旗旗,徐 敏.链路预测方法与网络结构的相关性[J].计算机技术与发展,2017,27(12):57.[doi:10.3969/ j. issn.1673-629X.2017.12.013]
　LI Qi-qi,XU Min.Correlation between Link Prediction Method and Network Structure[J].Computer Technology and Development,2017,27(03):57.[doi:10.3969/ j. issn.1673-629X.2017.12.013]
[4]陆圣宇,史 军,刘 宝,等.基于转移自洽和偏好连接的链路预测算法研究[J].计算机技术与发展,2019,29(08):18.[doi:10. 3969 / j. issn. 1673-629X. 2019. 08. 004]
　LU Sheng-yu,SHI Jun,LIU Bao,et al.Research on a Link Prediction Algorithm Based on Transferring Similarity and Preferential Attachment[J].Computer Technology and Development,2019,29(03):18.[doi:10. 3969 / j. issn. 1673-629X. 2019. 08. 004]
[5]初晓宇,高守玮.基于优先连接和用户属性的链路预测算法研究[J].计算机技术与发展,2019,29(11):17.[doi:10. 3969 / j. issn. 1673-629X. 2019. 11. 004]
　CHU Xiao-yu,GAO Shou-wei.Research on Link Prediction Algorithm Based on Professional Attachment and User Attributes[J].Computer Technology and Development,2019,29(03):17.[doi:10. 3969 / j. issn. 1673-629X. 2019. 11. 004]
[6]潘嘉琪,邹俊韬.一种基于高阶混合投影估计的网络嵌入方法[J].计算机技术与发展,2020,30(02):17.[doi:10. 3969 / j. issn. 1673-629X. 2020. 02. 004]
　PAN Jia-qi,ZOU Jun-tao.A Network Embedding Method Based on High-order Hybrid Projection Estimation[J].Computer Technology and Development,2020,30(03):17.[doi:10. 3969 / j. issn. 1673-629X. 2020. 02. 004]
[7]陈广福,江 玲,韩辉珍.基于节点度异质性惩罚的链路预测方法[J].计算机技术与发展,2022,32(12):81.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 013]
　CHEN Guang-fu,JIANG Ling,HAN Hui-zhen.Link Prediction Method Based on Node Degree Heterogeneity Penalization[J].Computer Technology and Development,2022,32(03):81.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 013]
[8]奚超亮,冷泳林.混合负采样的知识图谱嵌入[J].计算机技术与发展,2023,33(09):168.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 025]
　XI Chao-liang,LENG Yong-lin.Knowledge Graph Embedding with Mixed Negative Sampling[J].Computer Technology and Development,2023,33(03):168.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 025]

更新日期/Last Update: 2020-03-10