融合深度特征的Sword2vect商品在线评价情感分析

EMOTION ANALYSIS OF SWORD2VECT COMMODITY ONLINE EVALUATION WITH DEEP FEATURE

商品在线评论情感分析已经成为一个热门的研究话题。为了更好地解决情感分析中词语的上下文信息、深度信息和情感信息缺失问题,提出一种融合深度特征的Sword2vect情感分析方法。用情感得分加权word2vect词向量得到含有上下文和情感信息的Sword2vect;基于注意力的长短期记忆神经网络得到深度特征AttBilstm;融合深度特征的Sword2vect进行情感分析。在深度学习框架tensorflow进行实验并与已有的方法在准确率、召回率、F1等评价指标上进行比较,验证了该算法的有效性。

Emotion analysis of online reviews has become a hot topic of research.In order to better solve the context information of words in emotion analysis,the lack of depth information and emotional information,we propose a Sword2vect emotion analysis method with depth features.We used emotion score weighted word2vect word vector to get Sword2vect containing context and emotional information;the depth feature AttBilstm was obtained based on LSTM neural network of attention;depth features of Sword2vect were integrated for emotion analysis.The experiments are carried out in the deep learning framework tensorflow and compared with the existing methods under evaluation indexes of accuracy,recall,F1,the effectiveness of the algorithm is verified.