Chinese Term Extraction System Based on Mutual Information

An automatic Chinese term extraction system based on mutual information is presented.Firstly,the system gets term candidates by calculating internal associative strength of characters string using mutual information. Then, term selection from term candidates is done using basic word dictionary, common collocations suffix, prefix bank and some term POS composing rules. Experiment shows the precision is 72.19%, recall is 77.98% and F measure is 74.97%.