Spanish Emotion Recognition Method Based on Cross-Cultural Perspective.

Area of Science:

• Natural Language Processing
• Computational Linguistics
• Affective Computing

Background:

• Linguistic communication is crucial for cross-cultural understanding.
• Textual emotion recognition is vital for effective interpersonal communication.
• Spanish, as the second most spoken language globally, requires robust emotion analysis tools.

Purpose of the Study:

• To identify emotional features in Spanish texts.
• To develop an advanced framework for Spanish text sentiment classification.
• To enhance the accuracy of emotion recognition in Spanish language data.

Main Methods:

• Utilized an improved Bidirectional Long Short-Term Memory (BiLSTM) framework.
• Incorporated an attention mechanism to enhance feature extraction.
• Employed text preprocessing, emotion feature extraction, topic detection, and classification on Spanish dictionaries.

Main Results:

• Achieved an average emotion recognition accuracy of 76.21% on specialized Spanish dictionary datasets.
• The proposed BiLSTM framework with attention demonstrated superior performance.
• Outperformed comparable machine learning and convolutional neural network methods.

Conclusions:

• The improved BiLSTM framework with an attention mechanism is effective for Spanish text emotion recognition.
• The model shows enhanced generalization ability for recognizing emotion features.
• This approach offers a significant advancement in computational analysis of Spanish text sentiment.