基于word embedding和CNN的情感分类模型  被引量:20

Sentiment classification model based on word embedding and CNN

在线阅读下载全文

作  者:蔡慧苹 王丽丹[1] 段书凯[1] Cai Huiping;Wang Lidan;Duan Shukai(College of Electronic & Information Engineering, Southwest University, Chongqing 400715 , China)

机构地区:[1]西南大学电子信息工程学院,重庆400715

出  处:《计算机应用研究》2016年第10期2902-2905,2909,共5页Application Research of Computers

基  金:国家自然科学基金资助项目(61372139);国家教育部"春晖计划"科研资助项目(z2011148)

摘  要:尝试将word embedding和卷积神经网络(CNN)相结合来解决情感分类问题。首先,利用skip-gram模型训练出数据集中每个词的word embedding,然后将每条样本中出现的word embedding组合为二维特征矩阵作为卷积神经网络的输入,此外每次迭代训练过程中,输入特征也作为参数进行更新;其次,设计了一种具有三种不同大小卷积核的神经网络结构,从而完成多种局部抽象特征的自动提取过程。与传统机器学习方法相比,所提出的基于word embedding和CNN的情感分类模型成功地将分类正确率提升了5.04%。This paper tried to propose a method to solve the problem of sentiment classification by integrating word embeddingand convolutional neural network ( CNN) . First of all,the method accomplished a training process with skip-gram model to generateword embedding of each word in the dataset. Then,it created a two-dimensional feature matrix which was the combinationof word embedding of each word in a training sample as the input of CNN model. Each iteration process of training, entries offeature matrix would also update as part of model parameters. Secondly, this paper proposed a CNN structure which was mainlycomposed of three different sizes of convolution kernels so as to complete the automatic extraction process of a variety of localabstract features. Compared with traditional machine learning algorithms, the proposed word embedding and CNN based sentimentclassification model has successfully improved classification accuracy by 5. 04%.

关 键 词:卷积神经网络 自然语言处理 深度学习 词嵌入 情感分类 

分 类 号:TP183[自动化与计算机技术—控制理论与控制工程]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象