Skip to content

Latest commit

 

History

History
8 lines (5 loc) · 959 Bytes

README.md

File metadata and controls

8 lines (5 loc) · 959 Bytes

AECAT

Automatic Email Categorization_2009

Nowadays, emails are one of the most important information sources and ways to communicate. Emails are used for many and varied purposes generating a high amount of information.

This high amount of information requires a manipulation of emails intended to organize them on a specific taxonomy that makes it easier to find a particular message or a thread of messages on some subject; however, manual emails categorization is a time consuming task. Users’ satisfaction when using an email client might increase if the organization of emails according to specific interests may be achieved at a reduced effort.

This solution categorizes emails automatically while requiring a reduced effort from users. This solution is based on the d-Confidence active learning algorithm. We have achieved interesting results with low error rates at a reduced workload when compared to traditional fully supervised solutions.