<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0106">
  <Title>Making Hidden Markov Models More Transparent</Title>
  <Section position="4" start_page="34" end_page="34" type="metho">
    <SectionTitle>
3 Implementation
</SectionTitle>
    <Paragraph position="0"> The HMM part-of-speech tagging model and corresponding Viterbi algorithm were implemented based on their description in the updated version, http://www.cs.colorado.edu/~martin/ SLP/updated.html , of chapter 8 of (Jurafsky and Martin, 2000). A model was trained using Maximum Likelihood from the UPenn Treebank (Marcus et al., 1993). The input model file is encoded using XML and thus models built by other systems can be read in and displayed.</Paragraph>
    <Paragraph position="1"> The system is implemented in Java and requires 1.4 or higher to run. It has been tested on Linux and Apple operating systems. We will release it under a standard open source license.</Paragraph>
  </Section>
class="xml-element"></Paper>