Feature extraction and event/trend detection in social media messages

点击翻译按钮获取中文摘要

People are increasingly using social media (tweets, blogs and other forms of public posts) to express their opinions and concerns. Many companies are now interested in knowing what kind of chatter their brand is generating in the cyber space. To that effect, In the Chat has developed mining tools to retrieve and classify customer posts related to the products and services of specific companies. The proposed research will help In the Chat to improve its social media mining tools. More specifically we will develop new algorithms to automatically extract features from text messages that can facilitate their classification as well as new techniques to detect events or trends that may reflect changes in the public's opinion. This work will help In the Chat to maintain its leadership position in chat mining as well as Canadian and international retail companies to improve their customer service based on public feedback mined from social media.