<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-1056">
  <Title>Sydney, July 2006. c(c)2006 Association for Computational Linguistics Semi-Supervised Learning of Partial Cognates using Bilingual Bootstrapping</Title>
  <Section position="9" start_page="447" end_page="447" type="concl">
    <SectionTitle>
7 Conclusion and Future Work
</SectionTitle>
    <Paragraph position="0"> We showed that with simple methods and using available tools we can achieve good results in the task of partial cognate disambiguation.</Paragraph>
    <Paragraph position="1"> The accuracy might be increased by using dependencies relations, lemmatization, part-of-speech tagging - extract sentences where the partial cognate has the same POS, and other types of data representation combined with different semantic tools (e.g. decision lists, rule based systems). null In our experiments we use a machine language representation - binary feature values, and we show that nonetheless machines are capable of learning from new information, using an iterative approach, similar to the learning process of humans. New information was collected and extracted by classifiers when additional corpora were used for training.</Paragraph>
    <Paragraph position="2"> In addition to the applications that we mentioned in Section 1, partial cognates can also be useful in Computer-Assisted Language Learning (CALL) tools. Search engines for E-Learning can find useful a partial cognate annotator. A teacher that prepares a test to be integrated into a CALL tool can save time by using our methods to automatically disambiguate partial cognates, even though the automatic classifications need to be checked by the teacher.</Paragraph>
    <Paragraph position="3"> In future work we plan to try different representations of the data, to use knowledge of the relations that exists between the partial cognate and the context words, and to run experiments when we iterate the MB and BB steps more than once.</Paragraph>
  </Section>
class="xml-element"></Paper>