我国人工智能政策新词发现与演化研究OA北大核心CHSSCDCSSCICSTPCD

[Purpose/Significance]New word discovery,as the basic research of Chinese word segmentation,is very important for policy text mining,analysis and knowledge discovery.This paper intends to optimize the new word discovery algorithm and improve word segmentation performance in the AI policies.And we construct a domain lexicon to support the evolution study of AI policies.[Method/Process]A multi-feature fusion new word discovery algorithm(MFF)was pro-posed,which realized the mining of new word in the field of AI policies,and analyzed the evolution,continuation and dif-fusion of AI policies from the perspective of new word.[Result/Conclusion]The experimental results demonstrate the multi-feature fusion new word discovery algorithm effectively improves word segmentation performance and enriches the domain lexicon.The temporal changes in the emergence of AI policies neologisms reflect the key development areas of policy atten-tion at different stages,and reveal the characteristics and evolution process of central and local governments in terms of poli-cy innovation,continuation,diffusion and evolution.