| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To provide a method and a device for extracting a table from an unstructured sentence. <P>SOLUTION: This method for identifying a content table inside a document includes a step for generating a sequenced sequence of text fragments from the document and a step for selecting the content table as a continuous subsequence of the sequenced sequence of the text fragments when the content table satisfies the following criteria: each of entries defined by the text fragments of the content table has a link to a target text fragment having literal similarity with the entry, no target text fragment exists in the content table, and the target text fragment has ascending order corresponding to that of the entry defining the target text fragment. <P>COPYRIGHT: (C)2006,JPO&NCIPI</p> |
