<?xml version="1.0" standalone="yes"?>
<Paper uid="C00-1009">
  <Title>COMBINATION OF N-GRAMS AND STOCHASTIC CONTEXT-FREE GRAMMARS FOR LANGUAGE MODELING*</Title>
  <Section position="9" start_page="58" end_page="59" type="concl">
    <SectionTitle>
6 Conclusions
</SectionTitle>
    <Paragraph position="0"> A nc'w language model has been introduced.</Paragraph>
    <Paragraph position="1"> This new language model is detined as a~ linear ('olnl)in~ttion of an n-gram which repres(mts relations betwe('~n words, and a stochastic  the proposed language models in function of gamma. Different curves correspond to SCFGs estimated with different algorithms. The upper graphic correst)onds to the results obtained when the LRI algorithm was used in the language models, and the lower graphic corresponds to the results obtained with the VLRI algorithm.</Paragraph>
    <Paragraph position="2"> grammatical model which is used to represent the global relation between syntactic structures. The stochastic graminatical model is composed of a category-based SCFG and a probabilistic model of word distribution in the categories.</Paragraph>
    <Paragraph position="3"> Several algorithms have been described to estimate the parameters of the model flom a the smnple. In addition, efficient algorithms tbr solving the problem of the interpretation with this model have been presented.</Paragraph>
    <Paragraph position="4"> The proposed model has been tested on the part of Wall Street .Journal processed in the Penn Treebank project, and the results obtained improved by more tlmn 30% the test set; perplexity over results obtained by a simple 3-grain model.</Paragraph>
  </Section>
class="xml-element"></Paper>