Title :
Radical features for Chinese text classification
Author :
Hu, He ; Du, Xiaoyong
Abstract :
Chinese radicals play important roles in forming Chinese character´s semantic meaning. The semantic properties of radicals make them a promising source of information to be analyzed in text mining and content extraction. However, until recently there is little research work concentrating on using the radical set in text mining related tasks. We investigate the roles of radicals in Chinese text classification tasks. In the task, texts are transformed into vectors of radicals, characters and words. Radicals are further pruned by their semantic strengths and network traits. We carry out experiments with real data from Open Directory Project. The experiments results justify Chinese radicals as important features for semantic processing in Chinese text mining tasks.
Keywords :
natural language processing; pattern classification; text analysis; Chinese text classification; Open Directory Project; content extraction; radical features; radical set; semantic properties; text mining; Education; Europe; Positron emission tomography; Vectors; Chinese Radicals; Features; Text Classification;
Conference_Titel :
Fuzzy Systems and Knowledge Discovery (FSKD), 2012 9th International Conference on
Conference_Location :
Sichuan
Print_ISBN :
978-1-4673-0025-4
DOI :
10.1109/FSKD.2012.6234029