| Resumen |
Nowadays, people can write about their opinions of a product or service on social networks. These texts may contain information about a person’s likes or dislikes. Extracting sentiments or polarity from text is an important task to improve the quality or raise the popularity of a product/service or website. Unlike document and sentence analysis, aspect-based analysis is a more granular analysis that allows us to obtain specific information about the sentiment/polarity in a sentence or aspect. Various previous works have used transformer architectures, attention mechanisms, and neural networks with the purpose of obtaining new information between aspects and sentences. This article proposes a framework in which a pre-trained BERT model is used to obtain the word embeddings of the sentence and the aspect, then multi-label attention mechanisms are used fused with LSTM networks as a classifier, this framework can detect relevant information about the sentence and aspect. The results of our experiments show us that the proposed model obtains good results on the Laptop Restaurants data set. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025. |
