Marc Cheong

A microblogging-based approach to terrorism informatics: Exploration and chronicling civilian sentiment and response to terrorism events via Twitter

The study of terrorism informatics utilizing the Twitter microblogging service has not been given apt attention in the past few years. Twitter has been identified as both a potential facilitator and also a powerful deterrent to terrorism. Based on observations of Twitter’s role in civilian response during the recent 2009 Jakarta and Mumbai terrorist attacks, we propose a structured framework to harvest civilian sentiment and response on Twitter during terrorism scenarios. Coupled with intelligent data mining, visualization, and filtering methods, this data can be collated into a knowledge base that would be of great utility to decision-makers and the authorities for rapid response and monitoring during such scenarios. Using synthetic experimental data, we demonstrated that the proposed framework has yielded meaningful graphical visualizations of information, to reveal potential response to terrorist threats. The novelty of this study is that microblogging has never been studied in the domain of terrorism informatics. This paper also contributes to the understanding of the capability of conjoint structured data and unstructured content mining in extracting deep knowledge from noisy twitter messages, through our proposed structured framework.

Twittering for earth: a study on the impact of microblogging activism on earth hour 2009 in Australia

The role of Twitter - a form of microblogging - as both influencer and reflector of real-world events is fast emerging in todays world of Web 2.0 and social media. In this investigation, we survey how the use of Twitter in Australia is linked to the real-world success of the Earth Hour 2009 campaign. The results of this research will give us an idea of the emergence of microblogging as a new medium of influencing human behavior and providing a source of collective intelligence in planning and decision making, specifically in the Australian context. We found that, from our observations, there is a correlation between the inter-state total energy reduction during this campaign with the amount of interstate online Twitter discussion. We also identified a link between the Twitter discussion frequency and the total real-life population of the locale in which the chatter takes place, which could be used as a yardstick to analyze the reach of online technologies in the real world.

A Study on Detecting Patterns in Twitter Intra-topic User and Message Clustering

Timely detection of hidden patterns is the key for the analysis and estimating of driving determinants for mission critical decision making. This study applies Cheong and Lee’s “context-aware” content analysis framework to extract latent properties from Twitter messages (tweets). In addition, we incorporate an unsupervised Self-organizing Feature Map (SOM) as a machine learning-based clustering tool that has not been investigated in the context of opinion mining and sentimental analysis using microblogging. Our experimental results reveal the detection of interesting patterns for topics of interest which are latent and cannot be easily detected from the observed tweets without the aid of machine learning tools.

Dissecting Twitter: A Review on Current Microblogging Research and Lessons from Related Fields

Twitter as a microblogging service is fast gaining momentum in the past few years. Publications on the state of the art of various aspects Twitter are summarized in this review; which is structured to reflect the different categories of research that can be conducted on Twitter. The bulk of research has been identified to come from the message domain on Twitter, and so far progress has been made to bridge the two domains of Twitter (the message and the user), albeit in a limited form. This review also draws from findings in related fields that could be applied in the field of microblogging, such as research on the ‘blogosphere and blog trends; viral information on the web and memeties; and human factors in in-formation sharing and online presence, to name a few. This review provides re-searchers with an insight on to the various problem domains in microblogging re-search, and highlights the links between microblog research and other domains of research. Also, we show that current research rarely bridges the gap between the user and the message domains, and suggest potential improvements.

Twitmographics: Learning the Emergent Properties of the Twitter Community

This paper presents a framework for discovery of the emergent properties of users of the Twitter microblogging platform. The novelty of our methodology is the use of machine-learning methods to deduce user demographic information and online usage patterns and habits not readily apparent from the raw messages posted on Twitter. This is different from existing social network analysis performed on de facto social networks such as Face-book, in the sense that we use publicly available metadata from Twitter messages to explore the inherent characteristics about different segments of the Twitter community, in a simple yet effective manner. Our framework is coupled with the self-organizing map visualization method, and tested on a corpus of messages which deal with issues of socio politi-cal and economic impact, to gain insight into the properties of human interaction via Twitter as a medium for computer-mediated self-expression.

An approach to texture-based image recognition by deconstructing multispectral co-occurrence matrices using Tchebichef orthogonal polynomials

The existing use of summary statistics from co-occurrence matrices of images for texture recognition and classification has inadequacies when dealing with non-uniform and colored texture such as traditional `Batik¿ and `Songket¿ cloth motifs. This study uses the Tchebichef orthogonal polynomial as a way to preserve the shape information of cooccurrence matrices generated using the RGB multispectral method; allowing prominent features and shapes of the matrices to be preserved while discarding extraneous information. The decomposition of the six multispectral co-occurrence matrices yields a set of moment coefficients which can be used to quantify the difference between textures. The proposed method, when tested with a subset of the Vision Texture (VisTex) database and a collection of `Batik¿ and `Songket¿ motifs, yielded promising results of 99.5% and 95.28% classification rates respectively using the 3-nearest neighbor classifier.

Efficient textile recognition via decomposition of co-occurrence matrices

Textile motifs such as Batik and Songket are common native textile design throughout South East Asia, and are often imbued with cultural and spiritual meanings. However despite its cultural importance, automatic classification and retrieval work based on design motifs are not extensive. Previous work based on texture classification methods have proved successful but uses over 700 attributes. We show in this work that the number of attributes can be reduced down to 2% without significantly reducing the classification rate. This indicates that with the appropriate attribute reduction, fast recognition and classification of Batik and Songket textiles can be achieved.

Interpreting the 2011 London riots from twitter metadata

Social media have rapidly become one of the principal venues for personal and public communication. This makes them rich sources of information about real-world events. As a case study, we used Twitter metadata to investigate social dimensions of the 2011 London riots. The results showed that Twitter-based commentary and participation in the London Riots are closely linked to the real-world manifestation of the riots (e.g. in terms of geographic presence). Twitter metadata on users and their messages during the riots can be used to generate useful inferences which allows us to gain a better insight into intents, information-sharing behavior, and demographics of both the rioters and observers of the riots. Pattern recognition approaches can be used to further reveal latent properties from the acquired inferences.

Large-scale socio-demographic pattern discovery on microblog metadata

Microblogging services, such as Twitter, generate huge volumes of data reflecting the current zeitgeist. As such they are of enormous potential value to studies ranging from data mining to social anthropology. To realize the potential, this study investigates improvements of algorithms specifically tailored for the discovery of latent socio-demographic patterns in Twitter metadata. These newly improved hybrid algorithms improve on existing ones in terms of speed and scalability (from thousands of records to millions). Testing on a real-world Twitter data set (~7.4 million messages) reveals emergent patterns in global day-to-day Twitter activity. The results demonstrate novel insight when applied to real-world Twitter data, practical large-scale applications of the methods, and suggest potential areas of future research.