微博热度预测研究综述

doi:10.11871/jfdc.issn.2096-742X.2023.02.010

数据与计算发展前沿 ›› 2023, Vol. 5 ›› Issue (2): 119-135.

CSTR: 32002.14.jfdc.CN10-1649/TP.2023.02.010

doi: 10.11871/jfdc.issn.2096-742X.2023.02.010

微博热度预测研究综述

李妍^1,²(),何洪波^1,^*(),王闰强¹

1.中国科学院计算机网络信息中心，北京100083
2.中国科学院大学，北京100049

收稿日期:2022-02-17 出版日期:2023-04-20 发布日期:2023-04-24
通讯作者: 何洪波
作者简介:李妍，中国科学院计算机网络信息中心，硕士研究生，主要研究方向为新媒体技术应用、数据挖掘及应用。
本文中负责文献调研、文献分析、归纳总结与论文写作等。
LI Yan is a graduate student in the Compu-ter Network Information Center of Chinese Academy of Sciences. Her main research fields include the app-lication of new media technology, data mining, and applica-tion.
In this paper, she is responsible for literature research, literature analysis, summary, and thesis writing.
E-mail: yanli@cnic.cn|何洪波，中国科学院计算机网络信息中心，高级工程师，硕士生导师，主要研究方向为新媒体技术应用、互联网数据挖掘和信息推荐。
本文中负责写作指导、论文修订和总体统稿等。
HE Hongbo is a senior engineer and ma-ster tutor at the Computer Network Information Center of Chinese Academy of Sciences. His research fields include the application of new media technology, Internet data mining, and information recommendation.
In this paper, he is responsible for writing guidance, paper revision, and the final compilation.
E-mail: hhb@cnic.cn
基金资助:
中国科学院“十四五”网络安全和信息化专项子课题“网络空间科普云矩阵建设与应用”

A Survey of Research on Microblog Popularity Prediction

LI Yan^1,²(),HE Hongbo^1,^*(),WANG Runqiang¹

1. Computer Network Information Center, Chinese Academy of Sciences, Beijing 100083, China
2. University of Chinese Academy of Sciences, Beijing 100049, China

Received:2022-02-17 Online:2023-04-20 Published:2023-04-24
Contact: HE Hongbo

摘要/Abstract

摘要：

【目的】对现有微博热度预测研究展开多角度调研，讨论现有研究不足，展望未来发展趋势，为后续研究提供参考。【文献范围】本文整理和总结了近5年的国内外相关文献。【方法】本文首先介绍了热度预测问题的定义与热度计算方式，然后将热度预测研究方法从特征、时序和用户行为三个方面深入分析，再对热度预测问题的关键技术展开广泛调研，最后针对存在问题进行总结和展望。【结果】基于特征的热度预测方法因其定制性强被广泛使用，与深度学习和集成学习算法技术结合更是研究主流。【局限】由于各研究数据集未公开，本研究无法用统一的标准对所有算法技术的提升水平做横向对比。【结论】微博热度预测问题对于舆论监控、商业营销和内容推广等都具有一定意义，在社交媒体持续流行的时代，热度预测研究将会被继续深入推进。

关键词: 热度预测, 微博, 机器学习, 深度学习

Abstract:

[Objective] This paper is to conduct a multi-angle survey of the existing research on microblog popularity prediction, discuss the shortcomings of the existing approaches, foresee the future development trend, and provide a reference for follow-on researches. [Coverage] The paper sorts out and summarizes relevant literatures both in China and abroad in recent five years. [Methods] The paper first introduces the definition of popularity prediction and popularity calculation methods. Then the research methods of popularity prediction are analyzed from three aspects: characteristics, time sequence, and user behavior. An extensive study is conducted on the key technologies of popularity prediction. Finally, the problems of the existing methods and the prospect are summarized. [Results] Feature-based popularity prediction methods are widely used because of they are well customized. The method combining deep learning and ensemble learning is becoming the mainstream approach. [Limitations] As the dataset of the individual research is not publicly available, this study cannot make a horizontal comparison of all algorithms for the level of improvement against a unified standard. [Conclusions] Microblog popularity prediction is significant for public opinion monitoring, commercial marketing, and content promotion, etc. In the era of ever-increasing popularity of social media, the research on popularity prediction will be further promoted.

李妍,何洪波,王闰强. 微博热度预测研究综述[J]. 数据与计算发展前沿, 2023, 5(2): 119-135.

LI Yan,HE Hongbo,WANG Runqiang. A Survey of Research on Microblog Popularity Prediction[J]. Frontiers of Data and Computing, 2023, 5(2): 119-135, https://cstr.cn/32002.14.jfdc.CN10-1649/TP.2023.02.010.

图/表 6

表1

表2

表3

表4

热度预测常用评价指标-回归问题"

指标名称	英文简称	作用	公式
决定系数	R2	表示预测数据和真实数据之间拟合的程度	$1 - ∑ i = 1 m y i - y^i 2 ∑ i = 1 m y i - y - 2$
平均绝对误差	MAE	反映预测值误差的实际情况	$1 m ∑ i = 1 m y i - y^i$
平均绝对百分误差	MAPE	表示预测值和真实值之间平均偏差百分比	$1 m ∑ i = 1 m y i - y^i y i$
均方误差	MSE	用于评价数据的变化程度	$1 m ∑ i = 1 m y i - y^i 2$
均方根误差	RMSE	衡量预测值和真实值之间的偏差	$1 m ∑ i = 1 m y i - y^i 2$

表4

表5

热度预测常用评价指标-分类问题"

指标名称	英文简称	作用	公式
准确率	Accuracy	表示正确预测的数量占总样本的百分比	$A = (T P + T N) * 100 % T P + T N + F P + F N$
精确率	Precision	表示被正确地划分为正例的样本与预测结果为正例的样本数的比值	$P = T P * 100 % T P + F P$
召回率	Recall	表示有多少正例被判断为正例	$R = T P * 100 % T P + F N$
F1值	F1 Score	表示精确率和召回率的加权调和平均	$F 1 = 2 * R * P R + P$
受试者工作特征曲线	ROC	以FP Rate为横轴，TP Rate为纵轴的曲线	$F P R = F P F P + T N$ $T P R = T P T P + F N$
ROC曲线下面积	AUC	ROC曲线对应的面积	$A U C = ∑ i = 1 p R a n k i - M M + 1 2 M * N$

表5

表6

参考文献 84

[1]	肖显. 从抗击疫情看政务新媒体传播策略[J]. 青年记者, 2020, 80(30): 62-63.
[2]	Ji G, Zhu Y, Niu Y, et al. Classification and Evaluation for Microblog Popularity Prediction[C]. Journal of Phy-sics: Conference Series, IOP Publishing, 2021, 1883 (1): 012014.
[3]	Gao X, Cao Z, Li S, et al. Taxonomy and evaluation for microblog popularity prediction[J]. ACM Transactions on Knowledge Discovery from Data (TKDD), 2019, 13(2): 1-40.
[4]	Nia Z M, Khayyambashi M R. Improving content pop-ularity prediction with k-means clustering and deep-belief networks[J]. Multimedia Tools and Applications, 2021, 80(10): 15745-15764. doi: 10.1007/s11042-020-10463-x
[5]	新浪微博. 热门微博管理规范(试行版)[EB/OL]. [2014-10-17]. https://weibo.com/p/100160-376-671072-4380562.
[6]	王晓萌, 方滨兴, 张宏莉, 王星. TSL:基于连接强度的Facebook消息流行度预测模型[J]. 通信学报, 2019, 40(10):1-9. doi: 10.11959/j.issn.1000-436x.2019207
[7]	胡颖, 胡长军, 傅树深, 黄建一. 流行度演化分析与预测综述[J]. 电子与信息学报, 2017, 39(04):805-816.
[8]	郭庆光. 传播学教程(第1版)[M]. 北京: 中国人民大学出版社, 2005: 53.
[9]	刘定一, 沈阳阳, 詹天明, 刘亚军, 应毅. 融合微博热点分析和LSTM模型的网络舆情预测方法[J]. 江苏大学学报(自然科学版), 2021, 42(05):546-553.
[10]	冯新淇, 张琨, 任奕豪, 谢彬, 赵静. 一种基于RLDA主题模型的特征提取方法[J]. 计算机与数字工程, 2017, 45(10):1980-1985.
[11]	Gao X, Zheng Z, Chu Q, et al. Popularity prediction for single tweet based on heterogeneous bass model[J]. IEEE Transactions on Knowledge and Data Engineering, 2019, 33(5): 2165-2178.
[12]	刘钰. 政务新媒体用户的使用行为分析[D]. 天津大学, 2018.
[13]	清博智能. 清博指数榜单公式[EB/OL]. https://www.gsdata.cn/solution/zsmx.
[14]	Meng Q. Topic popularity prediction of online social network based on single objective evolution[J]. Inter-national Journal of Autonomous and Adaptive Commu-nications Systems, 2020, 13(4): 371-388.
[15]	李勇. 基于两层聚类的微博热点话题发现算法研究[J]. 自动化技术与应用, 2021, 40(11):45-50.
[16]	陈梦秋, 周安民. 基于SVM的新浪热门微博预测[J]. 现代计算机(专业版), 2017, 34(09):23-27.
[17]	郑志蕴, 江国林, 张行进, 王振飞, 李钝. 基于多特征的热门微博预测算法研究[J]. 小型微型计算机系统, 2017, 38(03):494-498.
[18]	于海, 吕晴晴, 时鹏, 王铮, 胡长军. 基于在线社交网络事件库多因素耦合的流行度预测方法[J]. 天津大学学报(自然科学与工程技术版), 2020, 53(12):1272-1280.
[19]	王新乐, 杨文峰, 廖华明, 王永庆, 刘悦, 俞晓明, 程学旗. 基于多维度特征的主题标签流行度预测[J]. 山东大学学报(理学版), 2020, 55(01):94-101.
[20]	Hoang T B N, Mothe J. Predicting information diffusion on Twitter-Analysis of predictive features[J]. Journal of computational science, 2018, 28(1): 257-264. doi: 10.1016/j.jocs.2017.10.010
[21]	Wang K, Bansal M, Frahm J M. Retweet wars: Tweet popularity prediction via dynamic multimodal regression[C]. 2018 IEEE Winter Conference on Applications of Computer Vision (WACV), IEEE, 2018: 1842-1851.
[22]	Zhu H, Yun X, Han Z. Weibo popularity prediction method based on propagation acceleration[J]. Journal of Computer Research and Development, 2018, 55(6): 1282-1293.
[23]	He Z, He Z, Wu J, et al. Feature construction for posts and users combined with LightGBM for social media popularity prediction[C]. Proceedings of the 27th ACM International Conference on Multimedia, 2019: 2672-2676.
[24]	Zhao J, Li J, Sui L, et al. Microblog Popularity Prediction Based on Multimodal Feature Fusion[C]. 2020 IEEE 3rd International Conference of Safe Production and Informatization (IICSPI), IEEE, 2020: 475-481.
[25]	Jia K, Zhang X. Micro-blog Retweeting Prediction Based on Combined-Features and Random Forest[C]. CCF Conference on Computer Supported Cooperative Work and Social Computing. Springer, Singapore, 2019: 429-440.
[26]	Lv J, Liu W, Zhang M, et al. Multi-feature fusion for pre-dicting social media popularity[C]. Proceedings of the 25th ACM international conference on Multimedia, 2017: 1883-1888.
[27]	Meghawat M, Yadav S, Mahata D, et al. A multimodal approach to predict social media popularity[C]. 2018 IEEE conference on multimedia information processing and retrieval (MIPR), IEEE, 2018: 190-195.
[28]	Gayberi M, Oguducu S G. Popularity prediction of posts in social networks based on user, post and image features[C]. Proceedings of the 11th International Conference on Management of Digital EcoSystems, 2019: 9-15.
[29]	Liu T, Zhong Y, Chen K. Interdisciplinary study on popularity prediction of social classified hot online events in China[J]. Telematics and Informatics, 2017, 34(3): 755-764.
[30]	Luo Y, Wang F, Zhao F, et al. A framework for policy information popularity prediction in new media[C]. 2019 IEEE International Conference on Intelligence and Security Informatics (ISI), IEEE, 2019: 209-211.
[31]	Cerqueira V, Torgo L, Smailović J, et al. A comparative study of performance estimation methods for time series forecasting[C]. 2017 IEEE international conference on data science and advanced analytics (DSAA), IEEE, 2017: 529-538.
[32]	Chen Xiaoliang, Lan Xiang, Wan Jihong, Lu Peng, Yang Ming, Pancioni Luca. Evolutionary Prediction of Nonst-ationary Event Popularity Dynamics of Weibo Social Network Using Time-Series Characteristics[J]. Discrete Dynamics in Nature and Society, 2021, 2021(1): 1-19.
[33]	Xiao C, Liu C, Ma Y, et al. Time sensitivity-based popu-larity prediction for online promotion on Twitter[J]. Information Sciences, 2020, 525(16): 82-92. doi: 10.1016/j.ins.2020.03.056
[34]	Xiao Y, Xie X, Li Q, et al. Nonlinear dynamics model for social popularity prediction based on multivariate chaotic time series[J]. Physica A: Statistical Mechanics and its Applications, 2019, 525(13): 1259-1275. doi: 10.1016/j.physa.2019.04.110
[35]	Bo Wu, Wen-Huang Cheng, Yongdong Zhang, Qiushi Huang, Jintao Li, and Tao Mei. Sequential prediction of social media popularity with deep temporal context networks[C]. InProceedings of the 26th International Joint Conference on Artificial Intelligence (IJCAI), 2017, 427(1):3062-3068.
[36]	Baosong M, Chenguang S. Prediction models for net-work multi-source dissemination of information based on multivariate chaotic time series[C]. 2017 3rd IEEE International Conference on Computer and Communi-cations (ICCC), IEEE, 2017: 767-771.
[37]	Sermsai R, Laohakiat S. Analysis and Prediction of Temporal Twitter Popularity Using Dynamic Time War-ping[C]. 2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE), IEEE, 2019: 176-180.
[38]	Song Y, Li A, Quan Y. Topics' popularity prediction based on ARMA model[C]. Proceedings of 2018 International Conference on Mathematics and Artificial Intelligence, 2018: 68-72.
[39]	Wu Q, Yang C, Zhang H, et al. Adversarial training model unifying feature driven and point process perspectives for event popularity prediction[C]. Proceedings of the 27th ACM International conference on information and knowledge management, 2018: 517-526.
[40]	张志扬, 张凤荔, 谭琪, 王瑞锦. 基于深度学习的信息级联预测方法综述[J]. 计算机科学, 2020, 47(07):141-153.
[41]	Wang H, Li Y, Feng Z, et al. ReTweeting analysis and prediction in microblogs: An epidemic inspired approach[J]. China Communications, 2013, 10(3): 13-24.
[42]	Cao Q, Shen H, Gao J, et al. Popularity prediction on social platforms with coupled graph neural networks[C]. Proceedings of the 13th International Conference on Web Search and Data Mining, 2020: 70-78.
[43]	Zhu H, Wang M. A Novel Reposting Prediction Method Based on Quantified Microblog Hotness in Sina Weibo[C]. Proc. International Conference on Computer Science and Application Engineering (Csae), 2017:13-18.
[44]	Quan Y, Jia Y, Zhou B, et al. Repost prediction incorp-orating time-sensitive mutual influence in social networks[J]. Journal of Computational Science, 2018, 28(5): 217-227. doi: 10.1016/j.jocs.2017.11.015
[45]	Li Y, Jin H, Yu X, et al. Intelligent prediction of private information diffusion in social networks[J]. Electronics, 2020, 9(5): 1-16. doi: 10.3390/electronics9010001
[46]	Xiao Y, Li J, Zhu Y, et al. User Behavior prediction of social hotspots based on multimessage interaction and neural network[J]. IEEE Transactions on Computational Social Systems, 2020, 7(2): 536-545. doi: 10.1109/TCSS.6570650
[47]	Chen F, Tan W H. Marked self-exciting point process modelling of information diffusion on Twitter[J]. The Annals of Applied Statistics, 2018, 12(4): 2175-2196.
[48]	Ma S, Feng L, Lai C H. Mechanistic modelling of viral spreading on empirical social network and popularity prediction[J]. Scientific reports, 2018, 8(1): 1-10.
[49]	Shang J, Huang S, Zhang D, et al. RNe2Vec: information diffusion popularity prediction based on repost network embedding[J]. Computing, 2021, 103(2): 271-289. doi: 10.1007/s00607-020-00858-x
[50]	Aghababaei S, Makrehchi M. Activity-based Twitter sampling for content-based and user-centric prediction models[J]. Human-centric Computing and Information Sciences, 2017, 7(1): 1-20. doi: 10.1186/s13673-016-0083-0
[51]	Han Z, Tang Z, He B. Improved Bass model for pre-dicting the popularity of product information posted on microblogs[J]. Technological Forecasting and Social Change, 2022, 176(3): 121458. doi: 10.1016/j.techfore.2021.121458
[52]	Cao Q, Shen H, Gao H, et al. Predicting the popularity of online content with group-specific models[C]. Proc-eedings of the 26th International Conference on World Wide Web Companion, 2017: 765-766.
[53]	Bao Z, Liu Y, Zhang Z, et al. Predicting popularity via a generative model with adaptive peeking window[J]. Physica A: Statistical Mechanics and its Applications, 2019, 522(10): 54-68. doi: 10.1016/j.physa.2019.01.132
[54]	Tan W H, Chen F. Predicting the popularity of tweets using internal and external knowledge: an empirical Bayes type approach[J]. AStA Advances in Statistical Analysis, 2021, 105(2): 335-352. doi: 10.1007/s10182-021-00390-z
[55]	Zhang H. Research on information popularity prediction of multimedia network based on fast K proximity alg-orithm[J]. International Journal of Autonomous and Adaptive Communications Systems, 2020, 13(2): 103-115. doi: 10.1504/IJAACS.2020.109808
[56]	Klubička F, Fernandez R. Examining a hate speech corpus for hate speech detection and popularity predi-ction[J]. arXiv preprint arXiv:1805.04661, 2018.
[57]	Zhang Y, Xu Z, Yang Q. Predicting popularity of mess-ages in twitter using a feature-weighted model[J]. .
[58]	Kristiyanti D A, Umam A H. Prediction of Indonesia presidential election results for the 2019-2024 period using twitter sentiment analysis[C]. 2019 5th Internat-ional Conference on New Media Studies (CONMEDIA), IEEE, 2019: 36-42.
[59]	Chen J, Sun B, Li H, et al. Deep ctr prediction in display advertising[C]. Proceedings of the 24th ACM intern-ational conference on Multimedia, 2016: 811-820.
[60]	武维, 李泽平, 杨华蔚, 林川, 王忠德. 融合内容特征和时序信息的深度注意力视频流行度预测模型[J]. 计算机应用, 2021, 41(07):1878-1884.
[61]	Li C, Ma J, Guo X, et al. Deepcas: An end-to-end pred-ictor of information cascades[C]. Proceedings of the 26th international conference on World Wide Web, 2017: 577-586.
[62]	Chen G, Kong Q, Xu N, et al. NPP: A neural popularity prediction model for social media content[J]. Neurocom-puting, 2019, 333(11): 221-230.
[63]	Yu H, Hu Y, Shi P. A prediction method of peak time popularity based on twitter hashtags[J]. IEEE Access, 2020, 8(1): 61453-61461. doi: 10.1109/Access.6287639
[64]	Zhang Y, Jatowt A. Image tweet popularity prediction with convolutional neural network[C]. European Confer-ence on Information Retrieval, Springer, Cham, 2019: 803-809.
[65]	Liao D, Xu J, Li G, et al. Popularity prediction on online articles with deep fusion of temporal process and content features[C]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33(01): 200-207.
[66]	Stokowiec W, Trzciński T, Wołk K, et al. Shallow reading with deep learning: Predicting popularity of online content using only its title[C]. International Symposium on Me-thodologies for Intelligent Systems, Springer, Cham, 2017: 136-145.
[67]	Liu Y, Zhao J, Xiao Y. C-RBFNN: A user retweet beh-avior prediction method for hotspot topics based on im-proved RBF neural network[J]. Neurocomputing, 2018, 275(5): 733-746. doi: 10.1016/j.neucom.2017.09.015
[68]	Zhang Z, Chen T, Zhou Z, et al. How to become instag-ram famous: Post popularity prediction with dualatte-ntion[C]. 2018 IEEE international conference on big data (big data), IEEE, 2018: 2383-2392.
[69]	李国成, 陆俊, 王赟, 黄瑞, 刘谋海. 基于Bagging二次加权集成的孤立森林窃电检测算法[J]. 电力系统自动化, 2022, 46(02):92-100.
[70]	Leo Breiman. Random Forests[J]. Machine Learning, 2001, 45(1) : 5-32. doi: 10.1023/A:1010933404324
[71]	Chen T Q, Guestrin C. XGBoost: a scalable tree boos-ting system[C]. Proceedings of the ACM SIGKDD Inter-national Conference on Knowledge Discovery and Data Mining, 2016: 785-794.
[72]	Qiu X, Zuo Y, Liu G. ETCF: An ensemble model for CTR prediction[C]. 2018 15th International Conference on Service Systems and Service Management (ICSSSM), IEEE, 2018: 1-5.
[73]	Chen J, Liang D, Zhu Z, et al. Social media popularity prediction based on visual-textual features with XGB-oost[C]. Proceedings of the 27th ACM International Conference on Multimedia, 2019: 2692-2696.
[74]	Tang J, Xu X, Qiu J. Model Construction and Evaluation of Microblog News Popularity Prediction[C]. 2021 7th Annual International Conference on Network and Infor-mation Systems for Computers (ICNISC), IEEE, 2021: 457-464.
[75]	Wang J, Lou C, Yu R, et al. Research on hot micro-blog forecast based on XGBOOST and random forest[C]. International Conference on Knowledge Science, Engine-ering and Management, Springer, Cham, 2018: 350-360.
[76]	Carta S, Podda A S, Recupero D R, et al. Popularity pred-iction of instagram posts[J]. Information, 2020, 11(9): 453-453. doi: 10.3390/info11090453
[77]	Kang P, Lin Z, Teng S, et al. Catboost-based framework with additional user information for social media popu-larity prediction[C]. Proceedings of the 27th ACM interna-tional conference on multimedia, 2019: 2677-2681.
[78]	Deshpande D. Prediction & evaluation of online news popularity using machine intelligence[C]. 2017 Intern-ational Conference on Computing, Communication, Con-trol and Automation (ICCUBEA), IEEE, 2017: 1-6.
[79]	Lin Z, Huang F, Li Y, et al. A layer-wise deep stacking model for social image popularity prediction[J]. World Wide Web, 2019, 22(4): 1639-1655. doi: 10.1007/s11280-018-0590-1
[80]	朱海龙, 云晓春, 韩志帅. 基于传播加速度的微博流行度预测方法[J]. 计算机研究与发展, 2018, 55(06):1282-1293.
[81]	Cao Q, Shen H, Cen K, et al. DeepHawkes: Bridging the gap between prediction and understanding of information cascades[C]. Proceedings of the 26th ACM International Conference on Information and Know-ledge Management, USA:ACM, 2017:1149-1158.
[82]	包恒彬, 马玉鹏, 杨奉毅, 韩云飞. 基于时空注意力机制的加油站级客流量预测[J]. 计算机工程, 2021, 47(04):291-297.
[83]	王萧萧, 王亭雯, 马玉玲, 范佳奕, 崔超然. 基于深度森林的P2P网贷借款人信用风险评估方法[J]. 计算机科学, 2021, 48(S2):429-434.
[84]	周志华. 机器学习(第1版)[M]. 北京: 清华大学出版社, 2016: 28-32.

特征类别	常见子特征
发博用户特征	认证、地区、标签、生日、注册时长、年龄、粉丝数、关注数、发博数、近期活跃度、过往被转评赞数、影响力等
博文内容特征	信息量、主题、情感倾向、语义、外链/表情/图片数量、是否包含图片/视频/链接/长微博/表情、是否原创、是否含话题标签、是否涉及明星/名人/影响力用户等
信息传播特征	传播层级、出入度、传播速度、传播子网体量大小等
时间特征	发布日期、发布时刻、时间差等
其他	图片/视频的视觉特征、话题所在领域背景等

文献编号	特征类别	关键子特征	有效性评估
[17]	发博用户特征	粉丝数()、过往转评赞数()	特征综合>发博用户特征>博文内容特征
[17]	博文内容特征	词频()、是否包含图片/视频/长微博()、是否包含表情(*)	特征综合>发博用户特征>博文内容特征
[20]	发博用户特征	粉丝数(*)、关注数()、过往点赞数()、所属组群数(*)	特征综合后有效
	博文内容特征	是否包含图片(*)
	时间特征	发布时间(*)
[25]	发博用户特征	粉丝数(*)、发博数()、过往转评赞数()、关注数()	特征综合>发博用户特征>博文内容特征
[25]	博文内容特征	微博平均长度(*)、情感词数量()、标签数()、@的用户数()	特征综合>发博用户特征>博文内容特征
[56]	发博用户特征	粉丝数()、关注数()、发博数(*)	特征综合>发博用户特征>博文内容特征
[56]	博文内容特征	是否被回复(***)	特征综合>发博用户特征>博文内容特征
[72]	发博用户特征	粉丝数()、关注数()	特征综合>发博用户特征>其他>博文内容特征
	博文内容特征	标签数(*)、概念()、标题()、子类别(*)
	其他	图像视觉特征(*)
[77]	发博用户特征	ID(*)、粉丝数()、过往总浏览量(**)	特征综合>发博用户特征>博文内容特征
[77]	博文内容特征	标签数(*)、概念()、图片数()、标题长度(*)	特征综合>发博用户特征>博文内容特征
[79]	发博用户特征	过往平均浏览量(**)、所属组群数()	特征综合>发博用户特征>博文内容特征>时间特征>其他
	博文内容特征	标签数()、文本长度()、标题长度(**)
	时间特征	发布时间(*)
	其他	图像视觉特征(*)

热度预测方法	特点	优势	劣势
基于特征的热度预测方法	静态、从博文视角出发、热度多为传播效果进入终态/稳定态时的热度	（1）研究成果较多，可参考性强；（2）针对性和定制性较强，可覆盖学者期望的所有特征。	（1）特征模型需要人工构建，耗时耗力，且特征与热度之间的相关性决定了模型的预测效果；（2）静态特征易受到社交网络中突发事件的冲击；（3）预测结果为热度的宏观数值，没有关注微观上个人的行为干预。
基于时序的热度预测方法	动态、从事件发展视角出发、体现的是传播过程的变化规律	（1）预测效果和精度普遍较好；（2）可针对未来某一时间节点进行预测，实用性强；（3）用过往观测数据预测未来时刻数据，可解释性较强。	（1）对于前期时序数据观测记录与整合工作要求较高；（2）观测时间区间本身的局限性会影响预测模型的效果；（3）对于未来时序的预测中，短时效果会优于长时效果。
基于用户行为的热度预测方法	从传播底层实现视角出发，强调用户之间的交互影响和传播行为	（1）具有较强的理论基础；（2）考虑用户行为的复杂性、动态性与多样性。	（1）该方法中如信息级联和传染病模型过于强调“邻居”带来的影响，极易陷入理想的理论模型中，与现实观测效果偏差较大；（2）传播规模较大时，底层行为网络的构建与计算会增大模型开销；（3）以传播范围的形式体现热度水平，应用场景有限。

微博热度预测研究综述

A Survey of Research on Microblog Popularity Prediction

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 6

参考文献 84

相关文章 15

编辑推荐

Metrics

本文评价

算法性质	文献编号	算法模型	研究特点	研究方法	任务类别	评价指标
传统机器学习	[56]	LR、SVM	聚焦特定情感倾向下的热度预测问题	特征	分类	ACC、F1
传统机器学习	[57]	FWM	引入特征加权机制	特征	分类	P、R、F1
深度学习	[35]	DTCN	将时间上下文和时间注意力结合	时序	回归	MAE、 SRC
	[39]	PreNets	创新的利用对抗模型寻找特征和时序点过程两种思维模式的平衡	特征、时序	回归	MAPE、Kendall
	[49]	RNe2Vec	可用于解决潜在用户关系网络未知的信息扩散热度预测问题	特征、用户行为	分类	ACC、P、R、F1
	[65]	DFTC	适用于事件传播任何阶段的热度预测	特征、时序	分类	ACC、F1
	[66]	BiLSTM	首次实现仅采用标题信息完成热度预测	特征	分类	ACC
集成学习	[72]	XGBoost	实现图片视觉特征、文本内容特征和用户特征的综合	特征	回归	MAE、 SRC
集成学习	[79]	LDS	用探测器根据特征性能自动设置集成模型深度	特征	回归	MAE、MSE

[1]	刘云帆,李琦,孙哲南,谭铁牛. 基于生成对抗网络的人脸年龄编辑方法综述[J]. 数据与计算发展前沿, 2023, 5(2): 2-23.
[2]	涂又友,郑奇靖,赵瑾. 基于深度学习方法研究分子/固体界面量子化质子耦合的电荷转移过程[J]. 数据与计算发展前沿, 2023, 5(2): 37-49.
[3]	高添,朱教君,张金鑫,孙一荣,于丰源,滕德雄,卢德亮,于立忠,王宗国. 基于新一代信息技术的温带森林生态系统碳通量精准计量[J]. 数据与计算发展前沿, 2023, 5(2): 60-72.
[4]	王凡,冯立强,曹荣强. 大数据驱动的海洋人工智能服务平台设计与应用[J]. 数据与计算发展前沿, 2023, 5(2): 73-85.
[5]	许淞源,刘峰. ESDRec：一种面向地球大数据平台的数据推荐模型[J]. 数据与计算发展前沿, 2023, 5(1): 55-64.
[6]	赵忠斌,蔡满春,芦天亮. 融合多头注意力机制的网络恶意流量检测[J]. 数据与计算发展前沿, 2022, 4(5): 60-67.
[7]	危婷,张宏海,蔺小丽,张蕾蕾,王妍,贾金峰. 云服务网站用户复访行为预测模型研究[J]. 数据与计算发展前沿, 2022, 4(3): 124-130.
[8]	孙永谦,张茹茹,林子涵,张圣林,谭智元,张玉志. KPI异常检测方法评估[J]. 数据与计算发展前沿, 2022, 4(3): 46-65.
[9]	陈琼,杨咏,黄天林,冯媛. 小样本图像语义分割综述[J]. 数据与计算发展前沿, 2021, 3(6): 17-34.
[10]	蒲晓蓉,黄佳欣,刘军池,孙家瑜,罗纪翔,赵越,陈柯成,任亚洲. 面向临床需求的CT图像降噪综述[J]. 数据与计算发展前沿, 2021, 3(6): 35-49.
[11]	何涛,王桂芳,马廷灿. 基于词嵌入语义异常的跨学科研究内容发现方法[J]. 数据与计算发展前沿, 2021, 3(6): 50-59.
[12]	张怡宁,何洪波,王闰强. 热门数字音频预测技术综述[J]. 数据与计算发展前沿, 2021, 3(4): 81-92.
[13]	蒲剑苏,朱正国,邵慧,高博洋,朱焱麟,闫宗楷,向勇. 基于可视化的固态电解质材料机器学习筛选与预测[J]. 数据与计算发展前沿, 2021, 3(4): 18-29.
[14]	张舒莹,韩鑫胤,何小雨,袁丹阳,栾海晶,李瑞琳,何佳茵,牛北方. 基于机器学习的基因组微卫星状态探测方法综述[J]. 数据与计算发展前沿, 2021, 3(3): 126-135.
[15]	陈子健,李俊,岳兆娟,赵泽方. 基于自编码器与属性信息的混合推荐模型[J]. 数据与计算发展前沿, 2021, 3(3): 148-155.