“Towards Re-Inventing Psychohistory”: Predicting the Popularity of Tomorrow’s News from Yesterday’s Twitter and News Feeds

Abstract

Rapid advances in machine learning combined with wide availability of online social media have created considerable research activity in predicting what might be the news of tomorrow based on an analysis of the past. In this work, we present a deep learning forecasting framework which is capable to predict tomorrow’s news topics on Twitter and news feeds based on yesterday’s content and topic-interaction features. The proposed framework starts by generating topics from words using word embeddings and K-means clustering. Then temporal topic-networks are constructed where two topics are linked if the same user has worked on both topics. Structural and dynamic metrics calculated from networks along with content features and past activity, are used as input of a long short-term memory (LSTM) model, which predicts the number of mentions of a specific topic on the subsequent day. Utilizing dependencies among topics, our experiments on two Twitter datasets and the HuffPost news dataset demonstrate that selecting a topic’s historical local neighbors in the topic-network as extra features greatly improves the prediction accuracy and outperforms existing baselines.

Keywords

Topic’s popularity / trend forecasting / social media

References

[1] AbbarS, CastilloC, SanfilippoA. To post or not to post: Using online trends to predict popularity of offline content. Proceedings of the 29th on Hypertext and Social Media, 2018215-219

[2] AhmedM, SpagnaS, HuiciF, NiccoliniS. A peek into the future: Predicting the evolution of popularity in user generated content. Proceedings of the 6th ACM International Conference on Web Search and Data Mining, 2013607-616

[3] AntonacciG, ColladonAF, StefaniniA, GloorP. It is rotating leaders who build the swarm: Social network determinants of growth for healthcare virtual communities of practice. Journal of Knowledge Management, 2017, 21(5): 1218-1239

[4] AralS V, AlstyneM. The diversity-bandwidth trade-off. American Journal of Sociology, 2011, 117(1): 90-171

[5] ArthurD, VassilvitskiiS. k-means++: The advantages of careful seeding. Proceedings of the 18th Annual ACM-SIAM Symposium on Discrete Algorithms, 20061027-1035

[6] BaccoucheM, MamaletF, WolfC, GarciaC, BaskurtA. Sequential deep learning for human action recognition. International Workshop on Human Behavior Understanding, 201129-39

[7] BandariR, AsurS, HubermanB A. The pulse of news in social media: Forecasting popularity. Sixth International AAAI Conference on Weblogs and Social Media, 2012

[8] BleiDM, NgAY, JordanMI. Latent dirichlet allocation. Journal of Machine Learning Research, 2003, 3: 993-1022

[9] CawkellAE. Science citation index. Effectiveness in locating articles in the anaesthetics field:“ perturbation of ion transport”. British Journal of Anaesthesia, 1971, 43(8): 814

[10] CovaB, CovaV. Tribal marketing. European Journal of Marketing, 2002, 36(5): 595-620

[11] AmorimRC, HennigC. Recovering the number of clusters in data sets with noise features using feature rescaling factors. Information Sciences, 2015, 324: 126-145

[12] ChoudhuryM, SundaramH, JohnA, SeligmannDD. Can blog communication dynamics be correlated with stock market activity?. Proceedings of the 19th ACM Conference on Hypertext and Hypermedia, 200855-60

[13] DieboldFX, MarianoRS. Comparing predictive accuracy. Journal of Business and Economic Statistics, 2002, 20(1): 134-144

[14] EbrahimiM, YazdavarAH, ShethA. Challenges of sentiment analysis for dynamic events. IEEE Intelligent Systems, 2017, 32(5): 70-75

[15] FreemanLC. A set of measures of centrality based on betweenness. Sociometry, 197735-41

[16] GilbertCHE, EricH. Vader: A parsimonious rulebased model for sentiment analysis of social media text. 8th International Conference on Weblogs and Social Media, 201482-91

[17] GloorPASociometrics and Human Relationships, 2017

[18] GloorPA, ColladonAF, de OliveiraJM, RovelliP, GalbierM, VogelM. Identifying tribes on twitter through shared context. Collaborative Innovation Networks, 201991-111

[19] GloorP, ColladonA d, OliveiraJM, RovelliP. Put your money where your mouth is: Using deep learning to identify consumer tribes from word usage. International Journal of Information Management, 2020, 51: 101924

[20] GravesA, MohamedAR, HintonG. Speech recognition with deep recurrent neural networks. 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, 20136645-6649

[21] GregorK, DanihelkaI, GravesA, RezendeD, WierstraD. DRAW: A Recurrent Neural Network for Image Generation. International Conference on Machine Learning, 20151462-1471

[22] GruhlD, GuhaR, KumarR, NovakJ, TomkinsA. The predictive power of online chatter. Proceedings of the 11th ACMSIGKDD International Conference on Knowledge Discovery in Data Mining, 200578-87

[23] GuptaRK, YangY. Predicting and understanding news social popularity withemotional salience features. Proceedings of the 27th ACM International Conference on Multimedia, 2019139-147

[24] HarveyD, LeybourneS, NewboldP. Testing the equality of prediction meansquared errors. International Journal of Forecasting, 1997, 13(2): 281-291

[25] HochreiterS, SchmidhuberJ. Long short-termmemory. Neural Computation, 1997, 9(8): 1735-1780

[26] HummonNP, DereianP. Connectivity in a citation network: The development of DNA theory. Social Networks, 1989, 11(1): 39-63

[27] KidaneYH, GloorPA. Correlating temporal communication patterns of the Eclipse open source community with performance and creativity. Computational and Mathematical Organization Theory, 2007, 13(1): 17-27

[28] KimSD, KimSH, ChoHG. Predicting the virtual temperature of web-blog articles as a measurement tool for online popularity. 2011 IEEE 11th International Conference on Computer and Information Technology, 2011449-454

[29] KleebR, GloorPA, NemotoK, HenningerM. Wikimaps: dynamic maps of knowledge. International Journal of Organisational Design and Engineering, 2012, 2(2): 204-224

[30] KraussJ, NannS, SimonD, GloorPA, FischbachK. Predicting Movie Success and Academy Awards through Sentiment and Social Network Analysis. 16th European Conference on Information Systems, 20082026-2037

[31] Misra Rishabh. News Category DatasetResearch-Gate, 2018

[32] NewmanMNetworks, 2018

[33] PenningtonJ, SocherR, ManningCD. Glove: Global vectors for word representation. Proceedings of the 2014 Conference on Empirical Methods in Natural Language processing, 20141532-1543

[34] PintoH, AlmeidaJM, GoncalvesMA. Using early view patterns to predict the popularity of YouTube videos. Proceedings of the 6th ACM International Conference on Web Search and Data Mining, 2013365-374

[35] RousseeuwPJ. Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics, 1987, 20: 53-65

[36] RuanY, PurohitH, FuhryD, ParthasarathyS, ShethAP. Prediction of Topic Volume on Twitter. Proceedings of the 4th International ACM Conference on Web Science, 2012397-402

[37] SzaboG, HubermanBA. Predicting the popularity of online content. Communications of the ACM, 2010, 53(8): 80-88

[38] TatarA, AntoniadisP D, AmorimMD, FdidaS. Ranking news articles based on popularity prediction. 2012 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, 2012106-110

[39] TatarA, LeguayJ, AntoniadisP, LimbourgA D, AmorimMD, FdidaS. Predicting the popularity of online articles based on user comments. Proceedings of the International Conference on Web Intelligence, Mining and Semantics, 20111-8

[40] WengL, MenczerF, AhnYY. Predicting successful memes using network and community structure. 8th International AAAI Conference on Weblogs and Social Media, 2014

[41] ZhangX, FuehresH, GloorPA. Predicting stock market indicators through twitter "I hope it is not as bad as I fear". Procedia-Social and Behavioural Sciences, 2011, 26: 55-62

RIGHTS & PERMISSIONS

Systems Engineering Society of China and Springer-Verlag GmbH Germany