Photo Credit: J. Semans |
IntroductionThe volume of short texts that have been generated every seconds via micro blogging platform like Twitter have become valuable information sources for various kind of activities. However, with the large amount, combined with the speed and dynamics the information is produced, it is is a great challenge for individuals or organizations to harness these information for their benefits. Our research team shall look at the problem of how information from this source can be made contextually relevant to personal needs, be it individual or organization. |
Project I: Personalization of Trending TweetsTweets personalization focuses on event-based tweets which can be tweeted by many parties. During occurence of events like #koreaferry, #mh370 etc., there are far too many tweets that are posted every second. The goal of tweets personalization is to show only tweets that are liked to be seen by a user. Scope
People
Deliverables Publications Lu Weilin, Gan Keng Hoon: Personalization of Trending Tweets Using Like-Dislike Category Model. Knowledge-Based and Intelligent Information & Engineering Systems 19th Annual Conference, KES-2015, Singapore, September 2015 Proceedings, Procedia Computer Science 60:236-245, Elsevier (2015). |
|
Project II: Multi Layers Classifiers for Short Texts ClassificationShort text classification poses a new challenge for the domain of text classification due to the nature of the texts which is short, sparse, diversity, as well as slang and informal language used. These characteristics create new challenges in text categorization as they often have negative effects on the classification performance. In this work, we focus on the development of a classification framework that addresses the mentioned issues, especially dealing with terms limitation in short texts like tweets. Scope
People
|
|