<?xml version="1.0" standalone="yes"?>
<Paper uid="P90-1039">
  <Title>A HARDWARE ALGORITHM FOR HIGH SPEED MORPHEME EXTRACTION AND ITS IMPLEMENTATION</Title>
  <Section position="8" start_page="312" end_page="312" type="evalu">
    <SectionTitle>
4.2 EVALUATION
</SectionTitle>
    <Paragraph position="0"> MEX-I works with 10MHz clock (i.e. the clock cycle is lOOns). Procedure 2, described in Section 3.1, including the simultaneous comparisons, is implemented for three clock cycles (i.e. 300ns).</Paragraph>
    <Paragraph position="1"> Then, the entire implementation time for morpheme extraction approximates A x D x L x M x 300n8. Here, D is the number of all morphemes in the dictionary, L is the length of input text, M is the number of text streams, and A is the indexing coef~dent. This coei~cient means the average rate for the number of compared morphemes, compared to the number of all morphemes in the dictionary.</Paragraph>
    <Section position="1" start_page="312" end_page="312" type="sub_section">
      <SectionTitle>
Results
</SectionTitle>
      <Paragraph position="0"> The implementation time measurement results, obtained for various kinds of Japanese text, are plotted in Fig. 10. The horizontal scale in Fig. 10 is the L x M value, which corresponds to the number of characters in all the text streams. The vertical scale is the measured implementation time.</Paragraph>
      <Paragraph position="1"> The above mentioned 80,000 morpheme dictionary was used in this measurement. These results show performance wherein MEX-I can extract morphemes from 10,000 character Japanese text by searching an 80,000 morpheme dictionary in 1 second.</Paragraph>
      <Paragraph position="2"> Figure 11 shows implementation time comparison with four conventional sequential algorithms. The conventional algorithms were carried out on NEC Personal Computer PC-98XL 2 (CPU: 80386, clock: 16MHz). Then, the 80,000 morpheme dictionary was on a memory board. Implementation time was measured for four diferent Japanese text samplings. Each of them forms one text stream, which includes 5,000 characters. In these measurement results, MEX-I runs approximately 1,000 times as fast as the morpheme extraction program, using the simple binary search algorithm.</Paragraph>
      <Paragraph position="3"> It runs approximately 100 times as fast as a program using the digital search algorithm, which has the highest speed among the four algorithms.</Paragraph>
      <Paragraph position="4">  toward achieving natural language parsing accelerators, which is a new approach to speeding up the parsing.</Paragraph>
      <Paragraph position="5"> The implementation time measurement results show performance wherein MEX-I can extract morphemes from 10,000 character Japanese text by searching an 80,000 morpheme dictionary in 1 second. When input is one stream of text, it runs 100-1,000 times faster than morpheme extraction programs on personal computers.</Paragraph>
      <Paragraph position="6"> It can treat multiple text streams, which are composed of character candidates, as well as one stream of text. The proposed algorithm is implemented on it in linear time for the number of candidates, while conventional sequential algorithms are implemented in combinational time. This is advantageous for character recognition or speech recognition.</Paragraph>
      <Paragraph position="7"> Its architecture is so simple that the authors believe it is suitable for VLSI implementation. Actually, its VLSI implementation is in progress. A high speed morpheme extraction VLSI will improve the performance of such text processing applications in practical use as Kana-to-Kanji conversion Japanese text input methods and spelling checkers on word processors, machine translation, automatic indexing for text database, text-to-speech conversion, and so on, because the morpheme extraction process is necessary for these applications.</Paragraph>
      <Paragraph position="8"> The development of various kinds of accelerator hardware for the other processes in parsing is work for the future. The authors believe that the hardware approach not only improves conventional parsing methods, but also enables new parsing methods to be designed.</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>