Analyzing the evolution of rare events via social media data and k-means clustering algorithm

Xiaoyu Sean Lu, Mengchu Zhou

Research output: Chapter in Book/Report/Conference proceedingConference contribution

15 Scopus citations

Abstract

Recently, many researchers attempt to find relationships between rare events and social media activities. This work proposes a data processing method based on the k-means clustering algorithm and analyze the evolution of a rare event via social media data. We use k-means twice in the spatial and time domains, respectively. The effectiveness of the method is verified by analyzing the damage of a hurricane named Sandy that occurred in 2012. The data set with respect to Sandy is obtained from a very popular social media, Twitter. The results show that our method can precisely predicate the accurate evolution of the hurricane, i.e., the affected place, time and severity. Besides, two new concepts, growth ratio and DRR rate, are presented to analyze the dataset in the time domain.

Original languageEnglish (US)
Title of host publicationICNSC 2016 - 13th IEEE International Conference on Networking, Sensing and Control
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781467399753
DOIs
StatePublished - May 25 2016
Externally publishedYes
Event13th IEEE International Conference on Networking, Sensing and Control, ICNSC 2016 - Mexico City, Mexico
Duration: Apr 28 2016Apr 30 2016

Publication series

NameICNSC 2016 - 13th IEEE International Conference on Networking, Sensing and Control

Other

Other13th IEEE International Conference on Networking, Sensing and Control, ICNSC 2016
Country/TerritoryMexico
CityMexico City
Period4/28/164/30/16

All Science Journal Classification (ASJC) codes

  • Artificial Intelligence
  • Computer Networks and Communications
  • Control and Systems Engineering
  • Modeling and Simulation

Keywords

  • Big data
  • data processing
  • k-means clustering
  • rare events

Fingerprint

Dive into the research topics of 'Analyzing the evolution of rare events via social media data and k-means clustering algorithm'. Together they form a unique fingerprint.

Cite this