AUTHOR=Liang Lin , Wang Shasha 

TITLE=Spanish Emotion Recognition Method Based on Cross-Cultural Perspective

JOURNAL=Frontiers in Psychology

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2022.849083

DOI=10.3389/fpsyg.2022.849083

ISSN=1664-1078

ABSTRACT=Linguistic communication is an important part of the cross-cultural perspective, and linguistic textual emotion recognition is a key piece of information in interpersonal communication. Spanish is the second largest language system in the world. This paper aims to identify the sentiment features in Spanish texts. The improved BiLSTM framework is proposed. We select three widely used Spanish dictionaries as the datasets for our experiments and then work through text pre-processing, text sentiment feature extraction, text topic detection, and sentiment classification to finally obtain text sentiment classification results. The attention mechanism is introduced in the improved BiLSTM framework. It enables the shared feature encoder to obtain weighted representation results in the process of extracting sentiment features, which enhances the model’s generalization capabilities for text sentiment feature recognition. Experimental results demonstrate that our approach performs better for specialized Spanish dictionary datasets. When it comes to sentiment recognition precision, the mean value reaches up to 76.21%. The overall performance outperforms current comparable machine learning methods and convolutional neural network methods.