DATE: | Thu, Jan 21, 2016 |
TIME: | 1:30 pm |
PLACE: | CBY A605 |
TITLE: | Identifying multiple topics in texts |
PRESENTER: |
Mohamed Mouine University of Ottawa |
ABSTRACT: Several methods exist to classify a text and identify the topic. These methods are mainly used to classify the news. But if the text covers several topic, the task becomes a bit more complicated. Our approach allows to analyze the text with more granularity to identify all topics. We rely on the text structure to cut it into several independent parts. We analyze thereafter each part separately. |