# Explainable Sentence-Level Sentiment Analysis for Amazon Product Reviews

**Authors:** Xuechun Li, Xueyao Sun, Zewei Xu, Yifan Zhou  
**arXiv:** [2111.06070](https://arxiv.org/abs/2111.06070)  
**Date:** 2021-11-11  
**Summary page:** https://www.summarizepaper.com/en/paper/2111.06070v1/

## Key points

- Authors Xuechun Li, Xueyao Sun, Zewei Xu, and Yifan Zhou conducted a study on sentence-level sentiment analysis for Amazon product reviews using a BiLSTM model with an attention mechanism.
- The researchers analyzed the distribution of attention weights within single sentences and main aspect terms.
- The BiLSTM model achieved an accuracy rate of up to 0.96, emphasizing the importance of considering both individual words and broader aspects in sentiment analysis.
- Main aspect terms in reviews received equal or greater attention weights compared to sentimental words in sentences, highlighting the significance of interpretability in text data analysis.
- This research provides valuable insights into sentiment analysis techniques and underscores the need for interpretability in this field, laying a foundation for future research.

## Summary

In their paper titled "Explainable Sentence-Level Sentiment Analysis for Amazon Product Reviews," authors Xuechun Li, Xueyao Sun, Zewei Xu, and Yifan Zhou delve into the realm of sentiment analysis by conducting a thorough examination of product reviews on Amazon. Utilizing a BiLSTM model with an attention mechanism, the researchers focus on sentence-level sentiment analysis and model interpretability. Their study includes an in-depth analysis of the distribution of attention weights within single sentences and main aspect terms. The results of their analysis reveal that the BiLSTM model achieves an impressive accuracy rate of up to 0.96. This finding highlights the importance of considering both individual words and broader aspects when analyzing sentiment in text data. The researchers also discover that main aspect terms within the reviews receive equal or even greater attention weights compared to sentimental words in sentences. This sheds light on the significance of interpretability in understanding how models process and analyze textual data. Overall, this research contributes valuable insights into the field of sentiment analysis and emphasizes the need for interpretability in this area. The findings presented in this paper offer a deeper understanding of sentiment analysis techniques and provide a foundation for future research in this area. By utilizing a BiLSTM model with an attention mechanism, the researchers were able to achieve an impressive accuracy rate of 0.96 in sentence-level sentiment analysis for Amazon product reviews. Through their thorough examination of attention weights within single sentences and main aspect terms, the authors reveal important insights into how models process and analyze text data. They also highlight the significance of considering not only individual words but also broader aspects when analyzing sentiment. This research serves as a valuable contribution to the field of sentiment analysis and emphasizes the need for interpretability in understanding how models process and analyze textual data. The results presented in this paper offer a deeper understanding of sentiment analysis techniques and provide a foundation for future research in this area.