Data mining of e-mail data on Java Platform

OData support
Supervisor:
Gáspár Csaba
Department of Telecommunications and Media Informatics

Using electronic mail have been being more and more popular recently. The companies and official electronic message exchange use email, despite the expansion of social networks. A user works as a manager can get up to hundreds of email a day, most of which should be in addition to also respond. It’s not just takes a lot of time to read these mails, buti t is very tiring. Therefore it is important to group emails and get prediction. It is a big advantage if we can be estimate when will the expected answer arrive. We can get interesting statistics if we combine e-mailing and data mining. In this way, users can get information about the habits and including an estimate for a given e-mail response duration.

The purpose of the thesis is to estimate the resply time from the saved suitable data, and then display the results for the user.

During my research, I studied literature, analyzed the existing solutions and possible technologies, and I have chosen the best option I considered, wich I realized the task with. Finally I evaluated the results.

I saved messages in the user database, which later I used in the RapidMiner software to build a model and then applied it. The results shows that the model can give a better estimate of the expected duration of the response, as if we would only examine the average user response time.

Downloads

Please sign in to download the files of this thesis.