基于对话结构与图注意力网络的药物推荐算法

doi:10.3778/j.issn.1673-9418.2309100

摘要/Abstract

摘要： 现有的药物推荐算法大多基于历史电子健康记录，但该数据难以反映患者的当前健康状况，也无法捕捉患者实时性的健康需求，导致推荐效果不佳。为此，融合了在线对话和疾病信息，提出一种基于对话结构与图注意力网络的药物推荐算法。集成灰关联分析与图注意力网络，运用灰关联分析学习节点间的关联，提出了一种新的关联感知图结构，以弥补传统图网络难以捕捉节点关联的不足。构建了对话分层编码器，基于新的图注意力网络编码话语与对话表示，并设计两种关联图结构学习节点的邻接关系，以生成蕴含上下文语义的对话结构表示。基于知识图谱和新的图网络学习疾病表示，将其与对话表示融合，实现药物的预测与推荐。实验结果表明，提出的算法在各评估方法下均优于基线方法，与性能最好的基线DNN相比，提出算法的F1和Jaccard分别提高了1.8%和3.5%，表明了提出算法能有效提高推荐性能。

关键词: 药物推荐, 对话结构, 图注意力网络, 灰关联分析

Abstract: Existing studies on medication recommendation are mostly based on the electronic health records. However, these data are difficult to reflect the current health status of patients and learn the current health needs of patients. Thus, this paper fuses the dialogue and disease information, and proposes a dialogue-based structure and graph attention network recommendation algorithm. Firstly, the grey relational analysis and graph attention network are integrated and the relations of nodes are measured using grey relational analysis to provide a novel relation-aware graph structure. It can improve the traditional graph networks to learn the node relations. Secondly, a dialogue hierarchical encoder is established to encode the representations of utterance and dialogue via the new graph structure. Meanwhile, two graph structures are designed to learn the node correlation for generating the contextual dialogue representation. Finally, the representations of disease by knowledge graph and graph network are incorporated with dialogue to predict and recommend. Experimental results show that the proposed algorithm is superior to baselines in terms of all metrics. Compared with the best baseline DNN, the performance of the proposed algorithm improves 1.8% and 3.5% in terms of F1 and Jaccard, which shows that the proposed algorithm can improve the recommendation performance.

Key words: medication recommendation, dialogue structure, graph attention network, grey relational analysis

陈江美, 张文德, 谭睿璞. 基于对话结构与图注意力网络的药物推荐算法[J]. 计算机科学与探索, 2024, 18(8): 2130-2139.

CHEN Jiangmei, ZHANG Wende, TAN Ruipu. Medication Recommendation Algorithm Based on Dialogue Structure and Graph Attention Network[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(8): 2130-2139.

参考文献

[1] ZHANG X, LIU S. Understanding relationship commitment and continuous knowledge sharing in online health communities: a social exchange perspective[J]. Journal of Knowledge Management, 2022, 26(3): 592-614.
[2] YANG H Z, GAO H Y. Personalized content recommendation in online health communities[J]. Industrial Management & Data Systems, 2021, 122(2): 345-364.
[3] FAN H M, LEDERMAN R. Online health communities: how do community members build the trust required to adopt information and form close relationships?[J]. European Journal of Information Systems, 2018, 27(1): 62-89.
[4] HU Y J, ZHOU S S. Will reviewer recommendation source and cured status bias review helpfulness in online health community?[J]. Online Information Review, 2023, 47(4): 680-696.
[5] HE Z F, HAN Y Q, OUYANG Z Q, et al. DialMed: a dataset for dialogue-based medication recommendation[C]//Proceedings of the 29th International Conference on Computational Linguistics, Oct 12-17, 2022: 721-733.
[6] CHOI E, BAHADORI M T, SUN J M, et al. Retain: an interpretable predictive model for healthcare using reverse time attention mechanism[C]//Advances in Neural Information Processing Systems 29, Barcelona, Dec 5-10, 2016: 29-37.
[7] WANG Y D, CHEN W T, PI D D, et al. Adversarially regularized medication recommendation model with multi-hop memory network[J]. Knowledge and Information Systems, 2021, 63: 125-142.
[8] SHANG J Y, XIAO C, MA T F, et al. GameNet: graph augmented memory networks for recommending medication combination[C]//Proceedings of the 33rd AAAI Conference on Artificial Intelligence, Jan 27-Feb 1, 2019. Menlo Park: AAAI, 2019: 1126-1133.
[9] YANG C Q, XIAO C, MA F L, et al. Safedrug: dual molecular graph encoders for recommending effective and safe drug combinations[C]//Proceedings of the 30th International Joint Conference on Artificial Intelligence, Aug 19-27, 2021: 3735-3741.
[10] REN Y J, SHI Y L, ZHANG K, et al. A drug recommendation model based on message propagation and DDI gating mechanism[J]. IEEE Journal of Biomedical and Health Informatics, 2022, 26(7): 3478-3485.
[11] ZHENG Z, QIU Z P, XIONG H, et al. DDR: dialogue based doctor recommendation for online medical service[C]//Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Aug 14-18, 2022. New York: ACM, 2022: 4592-4600.
[12] WEI Z Y, LIU Q L, PENG B L, et al. Task-oriented dialogue system for automatic diagnosis[C]//Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics, Jul 15-20, 2018. Stroudsburg: ACL, 2018: 201-207.
[13] YAN G J, PEI J H, REN P J, et al. ReMeDi: resources for multi-domain, multi-service, medical dialogues[C]//Procee-dings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval, Jul 11-15, 2022. New York: ACM, 2022: 3013-3024.
[14] BROCKI L, DYER G C, G?ADKA A, et al. Deep learning mental health dialogue system[C]//Proceedings of the 2023 IEEE International Conference on Big Data and Smart Computing, Feb 13-16, 2023. Piscataway: IEEE, 2023: 395-398.
[15] TIAN B, ZHANG Y, CHEN X H, et al. DRGAN: a GAN-based framework for doctor recommendation in Chinese on-line QA communities[C]//Proceedings of the 2019 International Conference on Database Systems for Advanced Applications, Apr 22-25, 2019: 444-447.
[16] 岳增营, 叶霞, 刘睿珩. 基于语言模型的预训练技术研究综述[J]. 中文信息学报, 2021, 35(9): 15-29.
YUE Z Y, YE X, LIU R H. A survey of language model based pre-training technology[J]. Journal of Chinese Information Processing, 2021, 35(9): 15-29.
[17] TORFI A, SHIRVANI R A, KENESHLOO Y, et al. Natural language processing advancements by deep learning: a survey[EB/OL]. [2023-07-15]. https://arxiv.org/abs/2003.01200.
[18] K?OSOWSKI P. Deep learning for natural language processing and language modelling[C]//Proceedings of the 2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications, Dec 6, 2018. Piscataway: IEEE, 2018: 223-228.
[19] VELI?KOVI? P, CUCURULL G, CASANOVA A, et al. Graph attention networks[EB/OL]. [2023-07-15]. https://arxiv.org/abs/1710.10903.
[20] LI Y C, YIN C X, ZHONG S H. Sentence constituent-aware aspect-category sentiment analysis with graph attention networks[C]//Proceedings of the 9th CCF International Conference on Natural Language Processing and Chinese Computing, Oct 14-18, 2019. Cham: Springer, 2020: 815-827.
[21] YUAN L, WANG J, YU L C, et al. Graph attention network with memory fusion for aspect-level sentiment analysis[C]//Proceedings of the 1st Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 10th International Joint Conference on Natural Language Processing, Dec 4-7, 2020. Stroudsburg: ACL, 2020: 27-36.
[22] ZHAO M, WANG L F, JIANG Z J, et al. Multi-task learning with graph attention networks for multi-domain task-oriented dialogue systems[J]. Knowledge-Based Systems, 2023, 259: 110069.
[23] WANG Y Z, WANG H Z, LU W B, et al. HyGGE: hyperbolic graph attention network for reasoning over knowledge graphs[J]. Information Sciences, 2023, 630: 190-205.
[24] 陈成, 张皞, 李永强, 等. 关系生成图注意力网络的知识图谱链接预测[J]. 浙江大学学报(工学版), 2022, 56(5): 1025-1034.
CHEN C, ZHANG H, LI Y Q, et al. Knowledge graph link prediction based on relational generative graph attention network[J]. Journal of Zhejiang University (Engineering Science), 2022, 56(5): 1025-1034.
[25] XU H, YUAN Z Q, ZHAO K, et al. GAR-Net: a graph attention reasoning network for conversation understanding[J]. Knowledge-Based Systems, 2022, 240: 108055.
[26] AHMED U, LIN J C W, SRIVASTAVA G. Hyper-graph attention based federated learning methods for use in mental health detection[J]. IEEE Journal of Biomedical and Health Informatics, 2022, 27(2): 768-777.
[27] HONG Y, LUO P Y, JIN S T, et al. LaGAT: link-aware graph attention network for drug-drug interaction prediction[J]. Bioinformatics, 2022, 38(24): 5406-5412.
[28] 邓聚龙. 灰色系统理论教程[M]. 武汉: 华中理工大学出版社, 1990.
DENG J L. Grey system theory course[M]. Wuhan: Huazhong University of Technology Press, 1990.
[29] 张岐山, 邓聚龙, 邵勇. 均衡接近度灰关联分析方法[J]. 华中理工大学学报, 1995, 23(11): 94-98.
ZHANG Q S, DENG J L, SHAO Y. A grey correlational analysis by the method of degree of balance and approach[J]. Journal of Huazhong University of Technology, 1995, 23(11): 94-98.
[30] 陈德光, 马金林, 马自萍, 等. 自然语言处理预训练技术综述[J]. 计算机科学与探索, 2021, 15(8): 1359-1389.
CHEN D G, MA J L, MA Z P, et al. Review of pre-training techniques for natural language processing[J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(8): 1359-1389.
[31] 王乃钰, 叶育鑫, 刘露, 等. 基于深度学习的语言模型研究进展[J]. 软件学报, 2021, 32(4): 1082-1115.
WANG N Y, YE Y X, LIU L, et al. Language models based on deep learning: a review[J]. Journal of Software, 2021, 32(4): 1082-1115.
[32] YE M C, LUO J Y, XIAO C, et al. LSAN: modeling long-term dependencies and short-term correlations with hierarchical attention for risk prediction[C]//Proceedings of the 29th ACM International Conference on Information & Know-ledge Management, Oct 19-23, 2020. New York: ACM, 2020: 1753-1762.
[33] LUO J Y, YE M C, XIAO C, et al. HiTANet: hierarchical time-aware attention networks for risk prediction on electronic health records[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Aug 20, 2020. New York: ACM, 2020: 647-656.
[34] SHEN W Z, WU S Y, YANG Y Y, et al. Directed acyclic graph network for conversational emotion recognition[EB/OL]. [2023-07-15]. https://arxiv.org/abs/2105.12907.