Link blog post to news articles
Detecting events from one or more temporally-ordered stream(s) of documents (e.g. news articles, blog posts) and group these documents based on the events that they describe is one of the goals in Topic Detection and Tracking (TDT). However, most of the existing event detection solutions do not co...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
2009
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/19306 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Detecting events from one or more temporally-ordered stream(s) of documents (e.g. news articles, blog posts) and group these documents based on the events that they
describe is one of the goals in Topic Detection and Tracking (TDT). However, most
of the existing event detection solutions do not consider users’ input (e.g. search
engines, blog posts) and group news articles into events which may not be useful to users.
In this project, the author studied the approach of query-guided event detection and
tracking from two parallel documents streams (news and blog) based on an ongoing research work. This approach takes users’ input into consideration through popular keyword queries and group queries, news articles and blog posts into events. The main focus in the project is to build an annotated dataset using real-world data collected from Google News and Technorati for evaluating the event detection algorithms. A web application was developed to facilitate the tasks of annotating, searching, analyzing and manipulating the dataset. Various software, tools and APIs were explored to aid in the development of a user friendly and interactive web interface. |
|---|