Convolutional Neural Networks Applied to Emotion Analysis in Texts: Experimentation from the Mexican Context

被引：1

作者：

Garduno-Miralrio, Juan-Carlos ^{[1
]}

Valle-Cruz, David ^{[1
]}

Lopez-Chau, Asdrubal ^{[1
]}

Rojas-Hernandez, Rafael ^{[1
]}

机构：

[1] Univ Autonoma Estado Mexico, Toluca, Mexico

来源：

KNOWLEDGE GRAPHS AND SEMANTIC WEB, KGSWC 2022 | 2022年 / 1686卷

关键词：

Convolutional neural network; Emotion analysis; Sentiment analysis; Accuracy; Social networks; Optimizers;

D O I：

10.1007/978-3-031-21422-6_10

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This work is twofold. First, it presents the state of the art of deep learning applied emotion analysis and sentiment analysis, highlighting the convolutional neural networks behavior over other techniques. Second, it presents experimentation on a convolutional neural network performance in the emotion analysis for the Mexican context, considering different architectures (with different number of neurons and different optimizers). The accuracy achieved in the proposed computational models is 0.9828 and 0.8943 with loss values of 0.1268 and 0.2387 respectively; however, the confusion matrices support the option of improving these models, giving the possibility of improving the values obtained and achieving greater accuracy.

引用

页码：133 / 148

页数：16