<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-1203">
  <Title>Measuring the Semantic Similarity of Texts</Title>
  <Section position="2" start_page="0" end_page="13" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Measures of text similarity have been used for a long time in applications in natural language processing and related areas. One of the earliest applications of text similarity is perhaps the vectorial model in information retrieval, where the document most relevant to an input query is determined by ranking documents in a collection in reversed order of their similarity to the given query (Salton and Lesk, 1971). Text similarity has been also used for relevance feedback and text classification (Rocchio, 1971), word sense disambiguation (Lesk, 1986), and more recently for extractive summarization (Salton et al., 1997b), and methods for automatic evaluation of machine translation (Papineni et al., 2002) or text summarization (Lin and Hovy, 2003).</Paragraph>
    <Paragraph position="1"> The typical approach to finding the similarity between two text segments is to use a simple lexical matching method, and produce a similarity score based on the number of lexical units that occur in both input segments. Improvements to this simple method have considered stemming, stop-word removal, part-of-speech tagging, longest subsequence matching, as well as various weighting and normalization factors (Salton et al., 1997a). While successful to a certain degree, these lexical matching similarity methods fail to identify the semantic similarity of texts. For instance, there is an obvious similarity between the text segments I own a dog and I have an animal, but most of the current text similarity metrics will fail in identifying any kind of connection between these texts. The only exception to this trend is perhaps the latent semantic analysis (LSA) method (Landauer et al., 1998), which represents an improvement over earlier attempts to use measures of semantic similarity for information retrieval (Voorhees, 1993), (Xu and Croft, 1996). LSA aims to find similar terms in large text collections, and measure similarity between texts by including these additional related words. However, to date LSA has not been used on a large scale, due to the complexity and computational cost associated with the algorithm, and perhaps also due to the &amp;quot;black-box&amp;quot; effect that does not allow for any deep insights into why some terms are selected as similar during the singular value decomposition process.</Paragraph>
    <Paragraph position="2"> In this paper, we explore a knowledge-based method for measuring the semantic similarity of texts. While there are several methods previously proposed for finding the semantic similarity of words, to our knowledge the application of these word-oriented methods to text similarity has not been yet explored. We introduce an algorithm  that combines the word-to-word similarity metrics into a text-to-text semantic similarity metric, and we show that this method outperforms the simpler lexical matching similarity approach, as measured in a paraphrase identification application.</Paragraph>
  </Section>
class="xml-element"></Paper>