CAJ | 학술논문

研究了Word2vec的工作原理及应用，明确了统计语言模型的关键问题，分析了词向量的特点，并对神经网络语言模型、Log＿Linear模型和Log＿Bilinear模型的基本原理进行了探讨，对Word2vec词向量训练框架的工作原理进行了详细分析，推导出了训练模型的目标函数，介绍了Word2vec工程的主要文件和训练参数，并将Word2vec应用于中文词向量的训练。
연구료Word2vec적공작원리급응용，명학료통계어언모형적관건문제，분석료사향량적특점，병대신경망락어언모형、Log＿Linear모형화Log＿Bilinear모형적기본원리진행료탐토，대Word2vec사향량훈련광가적공작원리진행료상세분석，추도출료훈련모형적목표함수，개소료Word2vec공정적주요문건화훈련삼수，병장Word2vec응용우중문사향량적훈련。
This paper studies the working principle and application of Word2vec,defines the key problems of statistical language model, analyzes the characteristics of word vector, probes into the basic principles of neural network language model,Log_Linear model and Log_Bilinear model,makes a detailed analysi on the working principle of word vector’s training framework of word2vec, and derives the objective functions of the training models,and introduces the main files in Word2vec project and training parameters, and applies Word2vec into the training of Chinese word vector.