Chinese Named Entity Extraction System Based On Word2vec Under Spark Platform
Jialu Yuan, Yongping Xiong · 2016
This paper proposes a real-time system that support the Chinese named entity extractions, which through word2vec algorithm training language mode to obtain word vector, and by calculating the Euclidean distance between word vectors to extract Chinese named entity, and transplant algorithm to Spark platform, using the Spark distributed computing ability improve training efficiency.First the system cut corpus into words with the help of existing word segmentation and get the rough corpus, then trains the rough corpus by word2vec algorithm to obtain word vectors and extracts the first layer of named entity according clustering algorithm, finally, the system uses the Named Entity Extraction(NEE) algorithm to extract the named entities and realize it on the spark platform.