<?xml version="1.0" standalone="yes"?>
<Paper uid="N06-2008">
  <Title>Temporal Classification of Text and Automatic Document Dating</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Various features have been used to classify and predict the characteristics of text and related text documents, ranging from simple word count models to sophisticated clustering and Bayesian models that can handle both linear and non-linear classes.</Paragraph>
    <Paragraph position="1"> The general goal of most classification research is to assign objects from a pre-defined domain (such as words or entire documents) to two or more classes/categories. Current and past research has largely focused on solving problems like tagging, sense disambiguation, sentiment classification, author and language identification and topic classification. We introduce an unsupervised method that classifies text and documents according to their predicted time of writing/creation. The method uses a sophisticated temporal language model to predict likely creation dates for a document, hence dating it automatically. This short paper presents some background information about existing techniques and the implemented system, followed by a brief explanation of the classification and dating method, and finally concluding with results and evaluation performed on the LDC</Paragraph>
  </Section>
class="xml-element"></Paper>