IEEE Computational Intelligence Magazine - August 2019 - 38

[31].These dense clusters were considered
social media events. In another work, a
system named DYNDENS was developed which quantified the magnitude of
change based on updates in edge weights.
The system incrementally computed
dense subgraphs to detect event stories
[32]. DYNDENS is efficient and scalable
to rapidly evolving datasets. Although the
detection methods proposed by [31], [32]
are efficient, despite rapid changes in
microblog streams, they suffer from the
loss of single-entity events.
Traditional event detection methods
are not designed to process and detect
events efficiently from such dynamic
data, particularly when the data stream is
noisy and consists of diverse events. In
addition, most of the state-of-the-art
approaches depend on highly weighted
and frequent patterns to detect events
[23], [25]-[27]. These approaches ignore
the dominating nature of burstiness over
small events in the data.
The proposed approach differs from
existing approaches because it highlights
dominating patterns at an early stage in
the text stream and handles post-event
effects by suppressing those patterns in
the subsequent time interval, which provides an opportunity to discover new
emerging events. Figure 5 visualizes the
pre-event, event, post-event graphs to
show the characteristics of the proposed
approach. Instead of focusing on burstiness, we considered change in temporal
frequency with respect to time which
we named displaced temporal frequency.
It captured the change in the frequencies of words appearing in text stream at
an early stage and later suppressed their
burstiness to highlight other topics.
These characteristics are an inherent
part of the proposed approach, which
lead to a better performance in the
event detection process.
5. Conclusion

In this paper, we presented a novel, sensitive and efficient Weighted Dynamic
Heartbeat Graph (WDHG) method to
detect events from a text stream. The
text stream was systematically transformed into a series of temporal graphs.
These graphs inherited temporal fre-

38

quencies and co-occurrence relationships of the words appearing in the text
stream. Each graph was further used to
extract a heartbeat score using two features: growth factor and aggregated centrality. A rule-based classifier labeled the
graphs as event candidates. Multiple
event candidates were merged together
to extract a list of ranked topics. For the
performance evaluation of the proposed
approach, three benchmarks: FA Cup,
Super Tuesday, and the US Election
were used. The quantitative evaluation
showed that the proposed approach outperformed the state-of-the-art methods.
The empirical evaluation showed that
the proposed approach is computationally efficient and scalable. In the future,
we plan to explore user participation
and social network based features, as
well as test the proposed approach on
live text streams.
References

[1] P. S. Earle, D. C. Bowden, and M. Guy, "Twitter
earthquake detection: Earthquake monitoring in a social
world," Ann. Geophys., vol. 54, no. 6, pp. 708-715, 2012.
[2] R. Ibrahim, A. Elbagoury, M. S. Kamel, and F. Karray, "Tools and approaches for topic detection from Twitter
streams: Survey," Knowl. Inf. Syst., vol. 54, no. 3, pp. 511-
539, 2018.
[3] M. A. Jarwar et al., "Communiments: A framework
for detecting community based sentiments for events,"
Int. J. Semantic Web Inf. Syst., vol. 13, no. 2, pp. 87-108,
2017.
[4] F. Johansson, V. Jethava, D. Dubhashi, and C. Bhattacharyya, "Global graph kernels using geometric embeddings," in Proc. 31st Int. Conf. Machine Learning, 2014,
pp. 1-9.
[5] F. D. Johansson and D. Dubhashi, "Learning with similarity functions on graphs using matchings of geometric embeddings," in Proc. 21st ACM SIGKDD Int. Conf. Knowledge
Discovery and Data Mining, ACM, 2015, pp. 467-476.
[6] E. Shabunina and G. Pasi, "A graph-based approach
to ememes identification and tracking in social media
streams," Knowl.-Based Syst., vol. 139, pp. 108-118,
2018.
[7] P. Yanardag and S. Vishwanathan, "Deep graph kernels," in Proc. 21st ACM SIGKDD Int. Conf. Knowledge
Discovery and Data Mining, ACM, 2015, pp. 1365-1374.
[8] P. Yanardag and S. Vishwanathan, "A structural
smoothing framework for robust graph comparison,"
in Proc. Advances in Neural Information Processing Systems,
2015, pp. 2134-2142.
[9] Y. Yao and L. B. Holder, "Detecting concept drift
in classification over streaming graphs," in Proc. KDD
Workshop on Mining and Learning with Graphs (MLG),
2016, pp. 2134-2142.
[10] Z. Saeed, R. A. Abbasi, M. I. Razzak, and G. Xu, "Event
detection in Twitter stream using weighted dynamic heartbeat graph approach," arXiv Preprint, arXiv:1902.08522,
2019.
[11] Z. Saeed, R. A. Abbasi, A. Sadaf, M. I. Razzak, and
G. Xu, "Text stream to temporal network: A dynamic
Heartbeat graph to detect emerging events on Twitter,"
in Proc. 22nd Pacific-Asia Conf. Advances in Knowledge Discovery and Data Mining, Springer, 2018, pp. 534-545.
[12] L. M. Aiello et al., "Sensing trending topics in Twitter," IEEE Trans. Multimedia, vol. 15, no. 6, pp. 1268-
1282, 2013.

IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE | AUGUST 2019

[13] M. Adedoyin-Olowe, M. M. Gaber, C. M. Dancausa,
F. Stahl, and J. B. Gomes, "A rule dynamics approach to
event detection in Twitter with its application to sports
and politics," Expert Syst. Appl., vol. 55, pp. 351-360,
2016.
[14] H.-J. Choi and C. H. Park, "Emerging topic detection in Twitter stream based on high utility pattern mining," Expert Syst. Appl., vol. 115, pp. 27-36, 2019.
[15] A. Elbagoury, R. Ibrahim, A. K. Farahat, M. S. Kamel, and F. Karray, "Exemplar-based topic detection in
Twitter streams," in Proc. 9th Int. AAAI Conf. Web and
Social Media, AAAI Publications, 2015, pp. 610-613.
[16] D. T. Nguyen and J. E. Jung, "Real-time event
detection for online behavioral analysis of big social
data," Future Gener. Comput. Syst., vol. 66, pp. 137-145,
2017.
[17] S. Papadopoulos, D. Corney, and L. M. Aiello,
"Snow 2014 data challenge: Assessing the performance
of news topic detection methods in social media," in Proc.
SNOW-DC@ WWW, 2014, pp. 1-8.
[18] R. Prabandari and H. Murfi, "Comparative study of
original recover and recover KL in separable non-negative matrix factorization for topic detection in Twitter,"
in Proc. AIP Conf., AIP, 2017, vol. 1862, p. 030144.
[19] Y. W. Teh, D. Newman, and M. Welling, "A collapsed variational Bayesian inference algorithm for latent Dir ich let a l location," in Proc. 19th Int. Conf.
Neural Information Processing Systems, MIT Press, 2007,
pp. 1353-1360.
[20] S. Petrovic´, M. Osborne, and V. Lavrenko, "Streaming first story detection with application to Twitter," in
Proc. Human Language Technologies: The 2010 Annu. Conf.
North American Chapter of the Association for Computational
Linguistics, Association for Computational Linguistics,
2010, pp. 181-189.
[21] K. Nur'aini, I. Najahaty, L. Hidayati, H. Murfi,
and S. Nurrohmah, "Combination of singular value decomposition and k-means clustering methods for topic
detection on Twitter," in Proc. 7th Int. Conf. Advanced
Computer Science and Information Systems, IEEE, 2015,
pp. 123-128.
[22] B. O'Connor, M. Krieger, and D. Ahn, "TweetMotif: Exploratory search and topic summarization for Twitter," in Proc. 4th Int. AAAI Conf. Weblogs and Social Media,
AAAI Publications, 2010, pp. 384-385.
[23] R. Li, K. H. Lei, R. Khadiwala, and K. C.-C. Chang,
"Tedas: A Twitter-based event detection and analysis system," in Proc. 28th Int. Conf. Data Engineering, IEEE, 2012,
pp. 1273-1276.
[24] M. Mathioudakis and N. Koudas, "Twittermonitor:
Trend detection over the Twitter stream," in Proc. 2010
ACM SIGMOD Int. Conf. Management of Data, ACM,
pp. 1155-1158.
[25] D. T. Nguyen and J. J. Jung, "Real-time event detection on social data stream," Mobile Netw. Appl., vol. 20,
no. 4, pp. 475-486, 2015.
[26] D. A. Shamma, L. Kennedy, and E. F. Churchill,
"Peaks and persistence: Modeling the shape of microblog
conversations," in Proc. 2011 ACM Conf. Computer Supported Cooperative Work, ACM, pp. 355-358.
[27] J. Yang and J. Leskovec, "Patterns of temporal variation in online media," in Proc. 4th ACM Int. Conf. Web
Search and Data Mining, ACM, 2011, pp. 177-186.
[28] Q. He, K. Chang, and E.-P. Lim, "Analyzing feature
trajectories for event detection," in Proc. 30th Annu. Int.
ACM SIGIR Conf. Research and Development in Information
Retrieval, ACM, 2007, pp. 207-214.
[29] J. Weng and B.-S. Lee, "Event detection in Twitter," in Proc. 5th Int. AAAI Conf. Weblogs and Social Media,
AAAI Publications, 2011, pp. 401-408.
[30] X. Cheng, X. Yan, Y. Lan, and J. Guo, "BTM: Topic
modeling over short texts," IEEE Trans. Knowl. Data
Eng., vol. 26, no. 12, pp. 2928-2941, 2014.
[31] M. K. Agarwal, K. Ramamritham, and M. Bhide,
"Real time discovery of dense clusters in highly dynamic
graphs: Identifying real world events in highly dynamic
environments," in Proc. VLDB Endowment, vol. 5, no. 10,
pp. 980-991, 2012.
[32] A. Angel, N. Sarkas, N. Koudas, and D. Srivastava,
"Dense subgraph maintenance under streaming edge
weight updates for real-time story identification," in Proc.
VLDB Endowment, vol. 5, no. 6, pp. 574-585, 2012.
IEEE Computational Intelligence Magazine - August 2019

Table of Contents for the Digital Edition of IEEE Computational Intelligence Magazine - August 2019
