NATIVE-SCRIPT AND CROSS-SCRIPT CHINESE NAME MATCHING

摘要

Techniques for Chinese name matching are described. A Chinese name is received and is romanized into a Mandarin Pinyin representation. The Mandarin Pinyin representation of the Chinese name is matched against a set of Romanized Chinese names originating from several different Chinese character names. In response to finding a potential match between the Mandarin Pinyin representation and Romanized Chinese name, the original Chinese script for the Romanized Chinese name is retrieved. A native script comparison is applied between the received Chinese name and the original Chinese script for the Romanized Chinese name to obtain a match score. The native script comparison includes character-by-character comparison, character variant look-up, and/or consideration of name component misalignments. The obtained match score is used as a filter to reduce false positives that are generated in the matching of the Mandarin Pinyin representation against the set of Romanized Chinese names.

主权项

1. A computer-implemented method for Chinese name matching, comprising:
receiving, by a processor, a Chinese name; romanizing, by the processor, the received Chinese name into a Mandarin Pinyin representation; matching, by the processor, the Mandarin Pinyin representation of the Chinese name against a set of Romanized Chinese names, wherein the Romanized Chinese names originate from a plurality of different Chinese character names; in response to finding a potential match between the Mandarin Pinyin representation and a Romanized Chinese name, retrieving, by the processor, the original Chinese script for the Romanized Chinese name; and applying, by the processor, a native script comparison between the received Chinese name and the original Chinese script for the Romanized Chinese name as a filter to reduce false positives that are generated in the matching of the Mandarin Pinyin representation of the Chinese name against the set of Romanized Chinese names.