Neologisms detection in a overlapping topical complex network

Neologism is a new word or new interpretation of an old word to describe and explain reality in a new way. As the root of knowledge, the detection work for neologism is essential but challenging. This paper presents a novel strategy for neologism detection from youngsters by an overlapping community detection method based on complex network which constructed by social texts. We pre-process our corpus of 600 thousands recorders from social texts (weibo.com) and model them with a topical complex network. And overlapping community detection work is realized with well-established algorithm EAGLE. For linguistics in social texts are productive and limited in dynamic pragmatics. We defined time coefficient and correlation coefficient and did some adaptive testing for the network construction. Experiments show that we can extract a topical network and obtain topic neologisms well in the real corpus. In the network, neologisms can be explained by topic words in the community of network.