Automatic Detection and Registration of Events by Analyzing Email Content

Abstract

In this paper we describe a system for automatic event detection and registration. This system monitors the received emails by the user and determines which of them are events based on their content. For analyzing the email content, Natural Language Processing techniques were used. Once an email is classified as an appointment, the details about this event are extracted and registered in the calendar of the user. Besides a mobile application notifies the user about this event and allows to view the specifics. 300 emails written in Spanish were collected to evaluate the performance of the system. A Naïve Bayes classifier was used to identify emails containing events and the specifics were extracted using pattern matching showing an overall system accuracy of 87%.