CHeM: A System for the Automatic Analysis of e-mails in the Restoration and Conservation Domain

Abstract

In this paper, we present the CHeM system, Cultural Heritage e-mail Manager, a support system for the analysis of e-mails of the Restoration and Conservation newsgroup, hosted by the Yahoo portal from December 2000 to January 2003. The complexity of the domain as well as the specificity of the e-mails, prompted us to build the first system prototype based on a client-side architecture, to help less expert users in classifying information contained in e-mails. The system goal is therefore to provide an instrument capable of classifying the received messages, downloaded onto the users' desktops, into standard categories, based on their content, using the well-known techniques of Data Mining and Information Retrieval. The categories thus obtained are then used to label the messages in order to provide valuable information on the domain and therefore support specific information retrieval and produce new user groups by an automatic generation of mailing lists. The methodology presented and the first test results are encouraging with a view to porting the system in other similar domains.

Resources

Details