A Corpus - Based Study Of Word News Headlines

Currently the study on network news headlines mainly using linguistic theory from some angles discuss language characteristic and the effect of expression.There is not enough on the quantitative.So the study which uses language corpus research and comparative analysis methods is few.This study on network news headlines based on linguistic theory use the methods of establishing a small corpus,made up of 854 thousand words. Then automatic word segmentation and POS tagging of corpus is used by special soft. A “network news headlines corpus” is finished.On this basis,with doing statistics the vocabulary frequency about the four website,we establish frequency table of network news headline and the vocabulary table of network news headline. The vocabulary of four websites before the 300 words are compared with the general vocabulary.Then we get the special vocabulary. In their special vocabulary of words,parole word mainly includes place names and national political figures.Language word includes the official title predicate and the common words of hot news events. Compared with high-frequency words and medium frequency words about the four websites,we find the similarities and differences in verbal words and language.In high frequency words,single syllable words are more,which are common words of news language.In high-frequency words and the second high-frequency words,the basic category words are more.In contrast,superior category words of the People’s Daily online and Xinhuanet are more than Sina and Netease network.In the medium-frequency words,the words about violence and pornography which come from Sina and Netease are more.Based on quantitative,the study combines with qualitative research methods. It develops into a “network news headline corpus”and completes “basic vocabulary table about network news headline”.It offer reference value for compiling news dictionaries and making news headline.