Identification and classification of proper nouns in Chinese texts

Various strategies are proposed to identify and classify three types of proper nouns in Chinese texts. Clues from character, sentence and paragraph levels are employed to resolve Chinese personal names. Character, Syllable and Frequency Conditions are presented to treat transliterated personal names. To deal with organization names, keywords, prefix, word association and parts-of-speech are applied. For fair evaluation, large scale test data are selected from six sections of a newspaper. The precision and the recall for these three types are (88.04%, 92.56%), (50.62%, 71.93%) and (61.79%, 54.50%), respectively. When the former two types are regarded as a category, the performance becomes (81.46%, 91.22%). Compared with other approaches, our approach has better performance and our classification is automatic.