<preface id="introduction"><title>Introduction</title><para><application>Select</application> is a tool for performing and evaluating email classification using supervised learning methods with incremental learning.After reading this section you should understand what that means.<!--, why it might be useful, and how this is done in the <application>Select</application>.--></para><sect1><title>The purpose of Select</title><para><application>Select</application> can be used for classification of text based documents, such as emails.Classification (also called categorization) is the task of assigning a class to a document.</para><para>(TODO Make an example)</para><sect2><title>Email classification</title><para>People who handle large quantities of email may benefit from using a classifier in the mail-handling process.</para><para>As a mail filter, a classifier could be used to automatically sort the incoming mail into one of a set of inboxes.This would allow the user to more easily prioritise which mail to read first.</para><para>As a mail filing assistant a classifier could be used to give suggestions to the user about which folder a message should be placed in.This could speed up the process of filing messages, because if the suggestions are correct the user would not have to search for the correct folder among the potentially large set of available folders.Even if the suggested folders are incorrect, the negative impact should be minimal, because then the user can simply file the message manually as usual.</para></sect2></sect1><sect1><title>An overview of Select</title><para><application>Select</application> follows a client/server model.There is a classification server, called <application>selectd</application>, that does all classification and learning.When a client wants to do some classification or learning then it asks the server to do it for them.</para><para>More to be written.</para></sect1></preface>