Determining semantic similarity of text based on its subsection

摘要

A system and method are provided for comparing a target sample of text with a set of text records.Each text record includes a text sample and one or more segments of the text in a text sample.The semantic similarity value between the target sample of the text and each of the text records is determined.Determining the specific semantic similarity between the target object of the text and the specific text record of the corpus is(I) determining individual semantic similarity values between the object samples of the text and each segment of the text indicated by a particular text recordAnd (II) based on individual semantic similarity valuesIncludes generating a specific semantic similarity between the target object of the text and the particular text record.The text record is then selected based on semantic similarity.