Social Media Understanding by Word Cloud Timeline

PACLIC 28 !2 Social Media Understanding by Word Cloud Timeline Virach Sornlertlamvanich Sirindhorn International Institute of Technology (SIIT), Thammasat University Pathum Thani 12121, Thailand virach@gmail.com Abstract Text from social media is significant key information to understand social movement. However, the length of the social media text is typically short and concise with a lot of absent words. Our task is to identify the proper keyword representing the message content that we are accounting for. Instead of training the model for keyword extraction directly from the Twitter messages, we propose a new method to fine-tune the model trained from some known documents containing richer context information. We conducted the experiment on Twitter messages and expressed in word cloud timeline. It shows a promising result. 1 Credits We adopted general Thai word segmentation module to extract the words and generate the key words for a specific domain based on the texts from Wikipedia1. The list of key words is then used to query the related tweets through the Twitter search API2 to collect the related tweets. In this study we propose an effective method to fine- tune the key words extracted from the document texts of Wikipedia to suit the relatively short texts from Twitter. The experiment and implementation have been conducted by Kobkrit Viriyayudhakorn. 1 http://th.wikipedia.org/ 2 https://dev.twitter.com/docs/streaming-api 2 Introduction Social media is a massive communication data for understanding the social behavior as well as sensing network is a massive monitoring data for observing the global environment (Gundecha and Liu, 2012). Both are the generated data that reflecting the real-time current situation of society and environment. In the rapid change of the current world, it is necessary to understand the situation and make a suitable response timely. The effect of happening or disaster nowadays has a trend to cause tremendous and pervasive damages. Since Great Hanshin earthquake in 1995, Indian Ocean earthquake and tsunami in 2004, Illinois hurricane Katrina in 2005, Arab spring a series of anti- government protests in 2011 uprising in Tunisia spread out to Yemen, Egypt, Syria, Libya and most of Arab countries, Tohoku earthquake and tsunami in 2011, Occupy Wall Street in 2011, until the recent Thailand coup d’etat in 2014, it is wondered whether we can learn something about these historical events. Focusing on social happenings, it is efficient enough to collect the social media data from the widely used social media applications such as Facebook, Twitter, Whatsapp, Line, or WeChat. Social media are actively used in most of the recent cases (Kaplan and Haenlein, 2010). If we ever view them in a proper dimension it is no doubt that we can somehow forecast, prevent, avoid the happenings by warning or influencing the communities to relief the disaster or the undesirable social situation development. In reality, social media data are vast, noisy, distributed, unstructured, and dynamic. Copyright 2014 by Virach Sornlertlamvanich 28th Pacific Asia Conference on Language, Information and Computation pages 2–4 PACLIC 28 !3 To study the evolution of social behavior on a happening, we …

Content truncated to 500 words