Online Community Conflict Decomposition with Pseudo Spatial Permutation
Online communities are composed of individuals sharing similar opinions or behavior in the virtual world. Facilitated by the fast development of social media platforms, the expansion of online communities have raised many attentions among the researchers, business analysts, and decision makers, leading to a growing list of literature studying the interactions especially conflicts in the online communities. A conflict is often initiated by one community which then attacks the other, leading to an adversarial relationship and worse social impacts. Many studies have examined the origins and process of online community conflict while failing to address the possible spatial effects in their models. In this paper, we explore the prediction of online community conflict by decomposing and analyzing its prediction error taking geography into accounts. Grounding on the previous natural language processing based model, we introduce pseudo spatial permutation to test the model expressiveness with geographical factors. Pseudo spatial permutation employs different geographical distributions to sample from and perturbs the model using the pseudo geographical information to examine the relationship between online community conflict and spatial distribution. Our analysis shows that the pseudo spatial permutation is an efficient approach to robustly test the conflict relation learned by the prediction model, and also reveals the necessity to incorporate geographical information into the prediction. In conclusion, this work provides a different aspect of analyzing the community conflict that does not solely rely on the textual communication.
KeywordsOnline community Spatial permutation Text mining Spatial social network Neural network
This material is partially based upon work supported by the National Science Foundation under Grant No. 1416509. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.
- 5.Kingma, D.P., Ba, J.: Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014)
- 6.Kumar, S., Hamilton, W.L., Leskovec, J., Jurafsky, D.: Community interaction and conflict on the web. In: Proceedings of the 2018 World Wide Web Conference on World Wide Web. pp. 933–943. International World Wide Web Conferences Steering Committee (2018)Google Scholar
- 9.Peters, M.E., et al.: Deep contextualized word representations. In: Proceedings of NAACL (2018)Google Scholar
- 10.Saxe, A.M., McClelland, J.L., Ganguli, S.: Exact solutions to the nonlinear dynamics of learning in deep linear neural networks. arXiv preprint arXiv:1312.6120 (2013)
- 12.Tang, D., Qin, B., Liu, T.: Document modeling with gated recurrent neural network for sentiment classification. In: Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pp. 1422–1432 (2015)Google Scholar
- 13.Wang, F., Lu, C.-T., Qu, Y., Yu, P.S.: Collective geographical embedding for geolocating social network users. In: Kim, J., Shim, K., Cao, L., Lee, J.-G., Lin, X., Moon, Y.-S. (eds.) PAKDD 2017. LNCS (LNAI), vol. 10234, pp. 599–611. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-57454-7_47CrossRefGoogle Scholar
- 19.Yue, Y., Dong, K., Zhao, X., Ye, X.: Assessing wild fire risk in the united states using social media data. J. Risk Res. 1–15 (2019). https://doi.org/10.1080/13669877.2019.1569098