基于情感信息融合注意力机制的抑郁症识别

doi:10.11959/j.issn.2096-6652.202221

智能科学与技术学报 ›› 2022, Vol. 4 ›› Issue (4): 600-609.doi: 10.11959/j.issn.2096-6652.202221

基于情感信息融合注意力机制的抑郁症识别

陈妍¹, 罗雪琴², 梁伟¹, 谢永芳³

¹ 湖南工商大学前沿交叉学院，湖南长沙 410205
² 湖南工商大学计算机学院，湖南长沙 410205
³ 中南大学自动化学院，湖南长沙 410083

修回日期:2021-12-17 出版日期:2022-12-15 发布日期:2022-12-01
作者简介:陈妍（1981− ），女，博士，湖南工商大学前沿交叉学院副教授，主要研究方向为大数据分析与智能决策、深度学习、智慧医疗
罗雪琴（1996− ），女，湖南工商大学计算机学院硕士生，主要研究方向为自然语言处理、深度学习、智慧医疗
梁伟（1982− ），男，博士，湖南工商大学前沿交叉学院副教授，主要研究方向为自然语言处理、大数据分析、深度学习、智慧医疗
谢永芳（1972− ），男，博士，中南大学自动化学院教授、博士生导师，主要研究方向为深度学习、智能优化、智能控制、知识自动化
基金资助:
国家自然科学基金资助项目(71971080);国家自然科学基金资助项目(71601077);湖南省自然科学基金资助项目(2020JJ4252)

Depression recognition based on emotional information fused with attentional mechanism

Yan CHEN¹, Xueqin LUO², Wei LIANG¹, Yongfang XIE³

¹ School of Advanced Interdisciplinary Studies, Hunan University of Technology and Business, Changsha 410205, China
² School of Computer Science, Hunan University of Technology and Business, Changsha 410205, China
³ School of Automation, Central South University, Changsha 410083, China

Revised:2021-12-17 Online:2022-12-15 Published:2022-12-01
Supported by:
The National Natural Science Foundation of China(71971080);The National Natural Science Foundation of China(71601077);The Natural Science Foundation of Hunan Province(2020JJ4252)

摘要/Abstract

摘要：

当前利用社交媒体数据预测抑郁症的研究忽视语言风格和情感随时间推移变化的特征，缺乏对情感状态和帖子元数据特征的研究，基于此，提出一种基于情感信息融合注意力机制的抑郁症识别模型。首先，在现有抑郁症识别研究的基础上，使用文本分类卷积神经网络按时间顺序提取用户每个时间段的帖子信息和情感信息，引入注意力机制为得到的特征矩阵分配不同的注意力权重，进而得到用户帖子信息特征和用户情感信息特征。其次，使用正则匹配提取文本的情感倾向信息，拼接注意力机制输出分配权重后的情感特征矩阵，以增强情感学习表示。然后，加入描述社交网络帖子的元数据特征，设计用于表示用户偏好特征的指标，通过统计表示指标提取用户语言偏好特征。最后，融合用户语言信息、用户情感信息、用户语言偏好信息3种特征，建立基于多层感知机的用户抑郁状态识别模型。实验结果显示，提出的模型精确率提高了0.051，召回率提高了0.065，F1值提高了0.058。

关键词: 抑郁症识别, 注意力机制, 情感信息, 深度学习

Abstract:

Aiming at the current research on using social media data to predict depression ignoring the characteristics of language style and emotional changes over time and lacking of research on the characteristics of emotional state and post metadata, a depression recognition model based on emotional information fused with attentional mechanism was proposed.Firstly, on the basis of the existing research on depression recognition, the text classification convolutional neural network was used to extract the post information and emotional information for each time period of the user in chronological order, and the attention mechanism was introduced to assign different attention weights to the obtained feature matrix to obtain user post information features and user emotional information features.Next, regular matching was used to extract the emotional tendency information, and the weighted emotional feature matrix output by the attention mechanism was spliced to enhance emotional learning expression.Then, metadata features describing social network posts were added, indicators that characterized user preference features were designed, and user language preference features through statistical characterization indicators were extracted.Finally, the three characteristics of user language information, user emotional information, and user language preference information were combined to establish a prediction model for the user’s depression state based on a multi-layer perceptron.The experimental results showed that the accuracy of the model in this paper had increased by 0.051, the recall rate had increased by 0.065, and F1 value had increased by 0.058.

Key words: depression recognition, attention mechanism, emotional information, deep learning

中图分类号:

陈妍, 罗雪琴, 梁伟, 等. 基于情感信息融合注意力机制的抑郁症识别[J]. 智能科学与技术学报, 2022, 4(4): 600-609.

Yan CHEN, Xueqin LUO, Wei LIANG, et al. Depression recognition based on emotional information fused with attentional mechanism[J]. Chinese Journal of Intelligent Science and Technology, 2022, 4(4): 600-609.

图/表 9

图1

图2

表1

表2

图3

图4

表3

表4

表5

参考文献 32

[1]	雍那, 杜莲, 胡华 ,等. 抑郁症患者攻击行为与应对方式、焦虑、抑郁的关系[J]. 第三军医大学学报, 2014,36(11): 1133-1137.
	YONG N , DU L , HU H ,et al. Correlation of aggressive behavior with coping styles,anxiety and depression in major depressed patients[J]. Journal of Third Military Medical University, 2014,36(11): 1133-1137.
[2]	CRUMP C , SUNDQUIST K , WINKLEBY M A ,et al. Mental disorders and vulnerability to homicidal death:Swedish nationwide cohort study[J]. BMJ (Clinical Research Ed), 2013,346:f557.
[3]	PARK M , CHA C , CHA M . Depressive moods of users portrayed in Twitter[J]. Proceedings of the ACM SIGKDD Workshop on Healthcare Informatics (HI-KDD), 2012: 1-8.
[4]	PARK M , MCDONALD D W , CHA M . Perception differences between the depressed and non-depressed users in twitter[C]// Proceedings of the 7th International AAAI Conference on Weblogs and Social Media.[S.l.:s.n.], 2013: 476-485.
[5]	XU R H , ZHANG Q P . Understanding online health groups for depression:social network and linguistic perspectives[J]. Journal of Medical Internet Research, 2016,18(3): e63.
[6]	DE CHOUDHURY M , GAMON M , COUNTS S ,et al. Predicting depression via social media[C]// Proceedings of the 7th International AAAI Conference on Weblogs and Social Media.[S.l.:s.n.], 2013: 128-137.
[7]	SHEN G Y , JIA J , NIE L Q ,et al. Depression detection via harvesting social media:a multimodal dictionary learning solution[C]// Proceedings of the 26th International Joint Conference on Artificial Intelligence. California:International Joint Conferences on Artificial Intelligence Organization, 2017: 3838-3844.
[8]	TROTZEK M , KOITKA S , FRIEDRICH C M . Utilizing neural networks and linguistic metadata for early detection of depression indications in text sequences[J]. IEEE Transactions on Knowledge and Data Engineering, 2018,32(3): 588-601.
[9]	KIM Y , . Convolutional neural networks for sentence classification[C]// Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing. Stroudsburg:Association for Computational Linguistics, 2014.
[10]	LIU Y , WU Y F . FNED:a deep network for fake news early detection on social media[J]. ACM Transactions on Information Systems, 2020,38(3): 1-33.
[11]	NASKAR D , SINGH S R , KUMAR D ,et al. Emotion dynamics of public opinions on twitter[J]. ACM Transactions on Information Systems, 2020,38(2): 1-24.
[12]	叶俊民, 罗达雄, 陈曙 . 基于短文本情感增强的在线学习者成绩预测方法[J]. 自动化学报, 2020,46(9): 1927-1940.
	YE J M , LUO D X , CHEN S . Short-text sentiment enhanced achievement prediction method for online learners[J]. Acta Automatica Sinica, 2020,46(9): 1927-1940.
[13]	BAHDANAU D , CHO K , BENGIO Y . Neural machine translation by jointly learning to align and translate[J]. arXiv preprint, 2014,arXiv:1409.0473.
[14]	沈继锋, 刘岳, 韦浩 ,等. 基于差分特征注意力机制的无锚框多光谱行人检测算法[J]. 智能科学与技术学报, 2021,3(3): 294-303.
	SHEN J F , LIU Y , WEI H ,et al. Anchor free multispectral pedestrian detection algorithm based on differential feature attention mechanism[J]. Chinese Journal of Intelligent Science and Technology, 2021,3(3): 294-303.
[15]	洪依, 孙成立, 冷严 . 基于超轻量通道注意力的端对端语音增强方法[J]. 智能科学与技术学报, 2021,3(3): 351-358.
	HONG Y , SUN C L , LENG Y . End-to-end speech enhancement based on ultra-lightweight channel attention[J]. Chinese Journal of Intelligent Science and Technology, 2021,3(3): 351-358.
[16]	宋婷, 陈战伟, 杨海峰 . 基于分层注意力网络的方面情感分析[J]. 大数据, 2020,6(5): 82-91.
	SONG T , CHEN Z W , YANG H F . Aspect sentiment analysis based on a hierarchical attention network[J]. Big Data Research, 2020,6(5): 82-91.
[17]	PENNINGTON J , SOCHER R , MANNING C . Glove:global vectors for word representation[C]// Proceedings of 2014 Conference on Empirical Methods in Natural Language Processing. Stroudsburg:Association for Computational Linguistics, 2014: 1532-1543.
[18]	VASWANI A , SHAZEER N , PARMAR N ,et al. Attention is all you need[J]. arXiv preprint, 2017,arXiv:1706.03762v5.
[19]	曾淦雄, 柯逍 . 基于 3D 卷积的图像序列特征提取与自注意力的车牌识别方法[J]. 智能科学与技术学报, 2021,3(3): 268-279.
	ZENG G X , KE X . 3D convolution-based image sequence feature extraction and self-attention for license plate recognition method[J]. Chinese Journal of Intelligent Science and Technology, 2021,3(3): 268-279.
[20]	HUTTO C , GILBERT E . VADER:a parsimonious rule-based model for sentiment analysis of social media text[C]// Proceedings of the 8th International AAAI Conference on Web and Social Media.[S.l.:s.n.], 2014.
[21]	LEIS A , RONZANO F , MAYER M A ,et al. Detecting signs of depression in tweets in Spanish:behavioral and linguistic analysis[J]. Journal of Medical Internet Research, 2019,21(6): e14199.
[22]	HUSSAIN J , SATTI F A , AFZAL M ,et al. Exploring the dominant features of social media for depression detection[J]. Journal of Information Science, 2020,46(6): 739-759.
[23]	CAVAZOS-REHG P A , KRAUSS M J , SOWLES S ,et al. A content analysis of depression-related Tweets[J]. Computers in Human Behavior, 2016,54: 351-357.
[24]	BENTON A , COPPERSMITH G , DREDZE M . Ethical research protocols for social media health research[C]// Proceedings of the 1st ACL Workshop on Ethics in Natural Language Processing. Stroudsburg:Association for Computational Linguistics, 2017: 94-102.
[25]	COPPERSMITH G , DREDZE M , HARMAN C . Quantifying mental health signals in twitter[C]// Proceedings of the Workshop on Computational Linguistics and Clinical Psychology:From Linguistic Signal to Clinical Reality. Stroudsburg:Association for Computational Linguistics, 2014: 51-60.
[26]	ZEILER M D . ADADELTA:an adaptive learning rate method[J]. arXiv preprint, 2012,arXiv:1212.5701.
[27]	SRIVASTAVA N , HINTON G E , KRIZHEVSKY A ,et al. Dropout:a simple way to prevent neural networks from overfitting[J]. Journal of Machine Learning Research, 2014,15(1): 1929-1958.
[28]	PEDREGOSA F , VAROQUAUX G , GRAMFORT A ,et al. Scikit-learn:machine learning in Python[J]. Journal of Machine Learning Research, 2012,12(10): 2825-2830.
[29]	SONG X M , NIE L Q , ZHANG L M ,et al. Multiple social network learning and its application in volunteerism tendency prediction[C]// Proceedings of the 38th International ACM SIGIR Conference on Research and Development in Information Retrieval.[S.l.:s.n.], 2015: 213-222.
[30]	ROLET A , CUTURI M , PEYRé G , . Fast dictionary learning with a smoothed Wasserstein loss[C]// Proceedings of the 18th Artificial Intelligence and Statistics.[S.l.:s.n.], 2016: 630-638.
[31]	CHUNG J , GULCEHRE C , CHO K ,et al. Empirical evaluation of gated recurrent neural networks on sequence modeling[J]. arXiv preprint, 2014,arXiv:1412.3555.
[32]	LIU N , LU P , ZHANG W ,et al. Knowledge-aware deep dual networks for text-based mortality prediction[C]// Proceedings of 2019 IEEE 35th International Conference on Data Engineering. Piscataway:IEEE Press, 2019: 1406-1417.

数据集	用户数/个	帖子数/条
Depressed	1 402	209 488
Non-Depressed	2 515	448 038

参数	值
lr	0.003
batchsize	20
hiddens	100
weight_decay	0.01
embed_size	200

预训练语料	精确率	召回率	F1值
Glove.200d	0.867	0.888	0.877
Glove.twitter.25d	0.845	0.847	0.846
Glove.twitter.50d	0.856	0.881	0.868
Glove.twitter.100d	0.859	0.895	0.877
Glove.twitter.200d	0.918	0.934	0.926

模型	精确率	召回率	F1值
NB	0.727	0.728	0.728
MSNL	0.818	0.818	0.818
WDL	0.769	0.768	0.768
MDL	0.848	0.850	0.849
GRU	0.825	0.823	0.824
TextCNN	0.825	0.852	0.838
BK-DDN	0.867	0.869	0.868
EIAM-DR	0.918	0.934	0.926

模型	精确率	召回率	F1值
EIAM-DR_(-sep)	0.853	0.875	0.864
EIAM-DR_(-ep)	0.869	0.904	0.886
EIAM-DR_(-p)	0.898	0.933	0.915
EIAM-DR	0.918	0.934	0.926

基于情感信息融合注意力机制的抑郁症识别

Depression recognition based on emotional information fused with attentional mechanism

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 32

相关文章 15

Metrics

推荐阅读 0

[1]	黄哲, 王永才, 李德英. 3D目标检测方法研究综述[J]. 智能科学与技术学报, 2023, 5(1): 7-31.
[2]	卢经纬, 程相, 王飞跃. 求解微分方程的人工智能与深度学习方法：现状及展望[J]. 智能科学与技术学报, 2022, 4(4): 461-476.
[3]	张俊, 许沛东, 陈思远, 高天露, 戴宇欣, 张科, 赵杭, 高杰迈, 白昱阳, 李金星, 张浩然, 李湘, 陈玖香. 物理-数据-知识混合驱动的人机混合增强智能系统管控方法[J]. 智能科学与技术学报, 2022, 4(4): 571-583.
[4]	栗仁武, 张凌霄, 高林, 李淳芃, 蒋浩. 基于点云的类级别物体姿态估计[J]. 智能科学与技术学报, 2022, 4(2): 246-254.
[5]	李琳辉, 周彬, 任威威, 连静. 行人轨迹预测方法综述[J]. 智能科学与技术学报, 2021, 3(4): 399-411.
[6]	沈继锋, 刘岳, 韦浩, 左欣, 杨万扣. 基于差分特征注意力机制的无锚框多光谱行人检测算法[J]. 智能科学与技术学报, 2021, 3(3): 294-303.
[7]	李颖, 陈龙, 黄钊宏, 孙杨, 蔡国榕. 基于多尺度卷积神经网络特征融合的植株叶片检测技术[J]. 智能科学与技术学报, 2021, 3(3): 304-311.
[8]	田庆, 胡蓉, 李佐勇, 蔡远征, 余兆钗. 基于SE-YOLOv5s的绝缘子检测[J]. 智能科学与技术学报, 2021, 3(3): 312-321.
[9]	尹开石, 杨萌, 顾曦, 王志成. 基于角度敏感的空间注意力机制的轻量型旋转目标检测器[J]. 智能科学与技术学报, 2021, 3(3): 322-333.
[10]	刘文, 胡琨林, 李岩, 刘钊. 移动目标轨迹预测方法研究综述[J]. 智能科学与技术学报, 2021, 3(2): 149-160.
[11]	张阳, 胡月, 辛东嵘. 一种考虑时空关联的深度学习短时交通流预测方法[J]. 智能科学与技术学报, 2021, 3(2): 172-178.
[12]	刘栋军, 王宇涵, 凌文芬, 彭勇, 孔万增. 基于脑机协同智能的情绪识别[J]. 智能科学与技术学报, 2021, 3(1): 65-75.
[13]	赵亮, 谢志峰, 张坤鹏, 郑玉卿, 付园坤. 无线网络信号传输建模：一种区间二型模糊集成深度学习方法[J]. 智能科学与技术学报, 2020, 2(4): 401-411.
[14]	田思佳,顾强,胡蓉,李锐戈,何顶新. 一种基于深度学习的机械臂分拣方法[J]. 智能科学与技术学报, 2020, 2(3): 268-274.
[15]	王飞跃,曹东璞,魏庆来. 强化学习：迈向知行合一的智能机制与算法[J]. 智能科学与技术学报, 2020, 2(2): 101-106.