【摘要】 为解决CNN和RNN无法同时兼顾局部和全局特征提取、传统情感分类模型不能充分利用原始评论文本中所蕴含的情感和语义信息、字词多义性调整、短文本语义稀疏等问题。提出一种基于ERNIE预训练模型和改进DPCNN的多通道的中文情感分类模型（MC＿E＿ImprovDPC＿B＿SV）来提高分类准确率，并实现准确分类。更多还原

【Abstract】 In order to solve the problems that CNN and RNN cannot simultaneously take into account the local and global feature extraction,the traditional emotion classification model cannot make full use of the emotional and semantic information contained in the original comment text,the polysemy adjustment of words,and the semantic sparseness of short texts.Therefore,a multi-channel Chinese sentiment classification model based on ERNIE pre-training model and improved DPCNN(MC＿E＿ImprovDPC＿B ＿SV)is proposed to improve the classification accuracy and achieve accurate classification.更多还原