Page 193 - Proceedings of the 2017 ITU Kaleidoscope

Challenges for a data-driven society

and Y.1712: OAM functionality for ATM-MPLS [3] V. Gupta, G. Lehal, “A Survey of Text Mining Techniques
interworking. From 2005 to 2008, PSTN (public switched and Applications,” Journal of Emerging Technologies In
telephone network) and ISDN (integrated service digital Web Intelligence, vol. 1, no. 1, pp. 60-76. 2009.
network) are extracted by TASIS. These keywords have a
correspondence relationship such as Recommendation ITU- [4] B. Liu, “Sentiment analysis and opinion mining,” Synthesis
T Y.2031: PSTN/ISDN emulation architecture. NGN (next Lectures on Human Language Technologies, vol. 5, no. 1,
generation network), IPv6 (Internet protocol version 6), and pp. 1-167, 2012.
IPTV (internet protocol television) are extracted by TASIS [5] S. Lee, S. Lee, H. Seol, and Y. Park, “Using patent information
from 2009 to 2012. NGN and IPv6 have a correspondence for designing new product and technology: keyword based
relationship such as NGN based on IPv6 [21]. Additionally, technology roadmapping,” R&D Management, vol. 38, pp.
NGN and IPTV have correspondence relationship such as 169-188, 2008.
Recommendations ITU-T Y.1900–Y.1999: IPTV over
NGN. Finally, Cloud and IoT are extracted by TASIS from [6] D. Blei, A. Ng, and M. Jordan, “Latent Dirichlet Allocation,”
2013 to 2016. These keywords have recently been described Journal of Machine Learning Research, vol. 3, pp. 993-1022,
as hot trend keywords. For example, the subject of 2003.
Recommendations ITU-T Y.3500–Y.3999 is Cloud
Computing and subject of Recommendations ITU-T [7] Z. Wu, N.E. Huang, S.R. Long, and C.K. Peng, “On the trend,
detrending, and variability of nonlinear and nonstationary time
Y.4000–4999 is Internet of Things and Smart Cities and series,” PNAS, vol. 104, no.38, pp. 14889-14894.
Communities. Therefore, TASIS shows that many
international standard documents including the cloud and [8] U. Flick, “An Introduction to Qualitative Research,” SAGE, vol.
IoT keywords have recently been published. In addition, 5, pp. 1-106.
Gartner Trends are also extracted these keywords [22].
[9] J.B. Michel, Y.K. Shen, A.P. Aiden, M.K. Gray, J.P. Pickett, D.
6. CONCLUSION Hoiberg, D. Clancy, P. Norvig, J. Orwant, S. Pinker, M.A.
Nowak, E.L. Aiden, The Google Books Team, “Quantitative
Analysis of Culture Using Millions of Digitized Books”,
We proposed TASIS, which automatically performs topic Science, vol. 331, pp.176-182.
modeling and trend analysis on document collections of the
ITU-T Recommendations. The TASIS based on an LDA [10] T.L. Griffiths, M. Steyvers, “Finding scientific topics,”
algorithm provides the results of topic modeling to users Proceedings of the National Academy of Sciences, vol. 101,
and a list of the documents relevant to each keyword in a no. 1, pp. 5528-5235, 2004.
topic. Moreover, TASIS also describes a TreeMap and trend
analysis graphs for easier understanding. [11] B. Levent, Ertekin. Seyda, and C. Lee Giles, “Topic and
Trend Detection in Text Collections Using Latent Dirichlet
Allocation,” ECIR, pp. 776-780, 2009.
Experiments with the Y series of ITU-T Recommendations
using TASIS have shown a limit in understanding the [12] M. Sergio, C. Paulo, and R. Paulo, “Business intelligence in
comprehensive trend analysis pattern. In future work, banking: A literature analysis from 2002 to 2013 using text
analysis of the entire documents in the ITU-T mining and latent dirichlet allocation,” Expert Systems with
Recommendations will be conducted. Additionally, research Applications, vol. 42, pp. 1314-1324, 2015.
to predict the publication of international standards for the
future technologies such as artificial intelligence or block [13] H. Thomas, “Probabilistic Latent Semantic Indexing,” SIGIR,
chain will be implemented. pp. 50-57, 1999.
[14] D. Blei, “Probabilistic Topic Models,” Communications of
ACKNOWLEDGEMENTS the ACM, vol. 55, pp. 77-84, 2012.

This work was supported by ICT R&D program of [15] Recommendation ITU-T Y Series, “Global information
MSIP/IITP. [2017-0-00455, Development of standards for infrastructure, Internet protocol aspects, next-generation
networks, Internet of Things and smart cities,” ITU-T
big data interoperability].
[16] Recommendation ITU-T Y.3501 (02/2016), “Cloud
REFERENCES computing – Framework and high-level requirements,” ITU-T,
pp. 1-28.
[1] M. Gary, E. John, H. Thomas, N. Robert, D. Dursun, and F.
Andrew, “Practical Text Mining and Statistical Analysis for [17] T.K. Moon, “The Expectation-Maximization Algorithm,”
Non-structured Text Data Applications,” Academic Press, pp. IEEE Signal Processing Magazine, vol. 13, pp. 47-60, 1996.-
3-126. 2012.
[18] Machine Learning for Language Toolkit, “Mallet API”,
[2] X. Hu, H. Liu, “Text analytics in social media, Mining text http://mallet.cs.umass.edu (visited on 2017-09-28).
data,” Springer US, pp. 385-414, 2012.
[19] Mike Bostock’s Blocks, “TreeMap API”,
https://bl.ocks.org/mbostock/4063582 (visited on 2017-09-28).

– 177 –