计算机应用与软件2018，Vol.35Issue(1)：97-102,174,7.DOI:10.3969/j.issn.1000-386x.2018.01.016
一种基于Word2Vec的训练效果优化策略研究
RESEARCH ON OPTIMIZATION STRATEGY OF TRAINING PERFORMANCE BASED ON WORD2VEC
摘要
Abstract
Word2Vec is a language process toolkit which was outsourced by Google in 2013.It can express words as real numbers based on neural network training language model,and find the words with high similarity in terms of vector cosine distance,and the training efficiency is higher.In this paper,we applied Word2Vec to find the optimized scheme in the task about training word vector with different segmented tools and mixed algorithms,meanwhile analysed architecture of the source code.Empirical results showed some factors and scheme which could improve the training performance,provided the higher quality word vectors for more applications.关键词
Word2Vec/词向量/语义相似度/算法Key words
Word2Vec/Word vector/Semantic similarity/Algorithm分类
信息技术与安全科学引用本文复制引用
王飞,谭新..一种基于Word2Vec的训练效果优化策略研究[J].计算机应用与软件,2018,35(1):97-102,174,7.基金项目
国家自然科学基金项目(11590771). （11590771）
