Quando a Nota Não Conta Toda a História: Inconsistências entre Ratings e Reviews em Sistemas de Recomendação
Resumo
Reviews and ratings are treated as complementary signals in recommender systems. However, users may assign numerical ratings that do not fully match the opinion expressed in their written reviews. This work analyzes this mismatch in the context of Review-Aware Recommender Systems (RARs), investigating whether textual reviews provide preference information beyond the original rating. We infer rating values from reviews with a zero-shot BERT-based sentiment model and compare them with the ratings assigned by users. We then evaluate the neighborhood structure induced by both signals and analyze the behavior of three RAR models when the original ratings are neutralized. The results show that rating-review inconsistencies occur in a non-negligible portion of the data, that review-inferred ratings produce higher neighborhood coverage in all datasets, and that the evaluated RARs remain highly sensitive to the original rating signal. These findings suggest that reviews contain relevant preference information that is not fully exploited by current review-aware recommendation models.
Palavras-chave:
Review-Aware Recommender Systems, Sentiment Analysis
Referências
Sumaia Al-Ghuribi, Shahrul Azman Mohd Noah, and Mawal Mohammed. 2023. An experimental study on the performance of collaborative filtering based on user reviews for large-scale datasets. PeerJ Comput Sci 9 (Aug. 2023), e1525.
Yan Andrade, Naan Silva, Adriano Machado Pereira, and Leonardo Rocha. 2026. Quando Comentários e Notas Discordam: Implicações para Sistemas de Recomendação Baseados em Reviews. In 41° Simpósio brasileiro de banco de dados (SBBD 2026).
Guilherme Bittencourt, Naan Vasconcelos, Yan Andrade, Nícollas Silva, Washington Cunha, Diego Roberto Colombo Dias, Marcos André Gonçalves, and Leonardo Rocha. 2026. Review-Aware Recommender Systems (RARSs): Recent Advances, Experimental Comparative Analysis, Discussions, and New Directions. ACM Comput. Surv. 58, 1, Article 21 (Sept. 2026), 49 pages. DOI: 10.1145/3744661
Cach Dang, María Moreno García, and Fernando De La Prieta. 2021. An Approach to Integrating Sentiment Analysis into Recommender Systems. Sensors 21 (08 2021), 5666. DOI: 10.3390/s21165666
Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), Jill Burstein, Christy Doran, and Thamar Solorio (Eds.). 4171–4186. DOI: 10.18653/v1/N19-1423
Gayatree Ganu, Noemie Elhadad, and Amélie Marian. 2009. Beyond the stars: improving rating predictions using review text content. In 12th International Workshop on the Web and Databases (WebDB). 1–6.
Dietmar Jannach, Markus Zanker, Alexander Felfernig, and Gerhard Friedrich. 2010. Recommender systems: an introduction. Cambridge University Press.
Chenliang Li, Cong Quan, Li Peng, Yunwei Qi, Yuming Deng, and Libing Wu. 2019. A Capsule Network for Recommendation and Explaining What You Like and Dislike. In Proceedings of the 42nd ACM SIGIR (SIGIR’19). DOI: 10.1145/3331184.3331216
Duantengchuan Li, Hai Liu, Zhaoli Zhang, Ke Lin, Shuai Fang, Zhifei Li, and Neal N Xiong. 2021. CARM: Confidence-aware recommender model via review representation learning and historical rating behavior in the online platforms. Neurocomputing 455 (2021), 283–296.
Hongtao Liu, Yian Wang, Qiyao Peng, Fangzhao Wu, Lin Gan, Lin Pan, and Pengfei Jiao. 2020. Hybrid neural recommendation with joint deep representation learning of ratings and reviews. Neurocomputing 374 (2020), 77–85.
Junrui Liu, Tong Li, Mingliang Yu, Shiqiu Yang, Zifang Tang, and Zhen Yang. 2025. A Multi-Factor Collaborative Prediction for Review-based Recommendation. In Proceedings of the Nineteenth ACM RecSys. DOI: 10.1145/3705328.3748062
Jianmo Ni, Jiacheng Li, and Julian McAuley. 2019. Justifying Recommendations using Distantly-Labeled Reviews and Fine-Grained Aspects. In Proceedings o EMNLP-IJCNLP, Kentaro Inui, Jing Jiang, Vincent Ng, and Xiaojun Wan (Eds.). DOI: 10.18653/v1/D19-1018
Francesco Ricci, Lior Rokach, and Bracha Shapira. 2015. Recommender systems: introduction and challenges. Recommender systems handbook (2015), 1–34.
Md Shohel Sayeed, Varsha Roji, and Kalaiarasi Anbananthen. 2023. BERT: A Review of Applications in Sentiment Analysis. HighTech and Innovation Journal 4 (06 2023), 453–462. DOI: 10.28991/HIJ-2023-04-02-015
Mehdi Srifi, Ahmed Oussous, Ayoub Ait Lahcen, and Salma Mouline. 2020. Recommender systems based on collaborative filtering using review texts—a survey. Information 11, 6 (2020), 317.
NLP Town. 2019. bert-base-multilingual-uncased-sentiment. [link]
Heitor Werneck, Nícollas Silva, Matheus Carvalho Viana, Fernando Mourão, Adriano CM Pereira, and Leonardo Rocha. 2020. A survey on point-of-interest recommendation in location-based social networks. In Proceedings of the Brazilian Symposium on Multimedia and the Web. 185–192.
Frank Wilcoxon. 1992. Individual Comparisons by Ranking Methods. Springer New York, New York, NY, 196–202. DOI: 10.1007/978-1-4612-4380-9_16
Huiqiong Wu, Guibing Guo, Enneng Yang, Yudong Luo, Yadan Chu, Linying Jiang, and XingweiWang. 2024. PESI: Personalized Explanation recommendation with Sentiment Inconsistency between ratings and reviews. Knowledge-Based Systems 283 (2024), 111133.
Yan Andrade, Naan Silva, Adriano Machado Pereira, and Leonardo Rocha. 2026. Quando Comentários e Notas Discordam: Implicações para Sistemas de Recomendação Baseados em Reviews. In 41° Simpósio brasileiro de banco de dados (SBBD 2026).
Guilherme Bittencourt, Naan Vasconcelos, Yan Andrade, Nícollas Silva, Washington Cunha, Diego Roberto Colombo Dias, Marcos André Gonçalves, and Leonardo Rocha. 2026. Review-Aware Recommender Systems (RARSs): Recent Advances, Experimental Comparative Analysis, Discussions, and New Directions. ACM Comput. Surv. 58, 1, Article 21 (Sept. 2026), 49 pages. DOI: 10.1145/3744661
Cach Dang, María Moreno García, and Fernando De La Prieta. 2021. An Approach to Integrating Sentiment Analysis into Recommender Systems. Sensors 21 (08 2021), 5666. DOI: 10.3390/s21165666
Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), Jill Burstein, Christy Doran, and Thamar Solorio (Eds.). 4171–4186. DOI: 10.18653/v1/N19-1423
Gayatree Ganu, Noemie Elhadad, and Amélie Marian. 2009. Beyond the stars: improving rating predictions using review text content. In 12th International Workshop on the Web and Databases (WebDB). 1–6.
Dietmar Jannach, Markus Zanker, Alexander Felfernig, and Gerhard Friedrich. 2010. Recommender systems: an introduction. Cambridge University Press.
Chenliang Li, Cong Quan, Li Peng, Yunwei Qi, Yuming Deng, and Libing Wu. 2019. A Capsule Network for Recommendation and Explaining What You Like and Dislike. In Proceedings of the 42nd ACM SIGIR (SIGIR’19). DOI: 10.1145/3331184.3331216
Duantengchuan Li, Hai Liu, Zhaoli Zhang, Ke Lin, Shuai Fang, Zhifei Li, and Neal N Xiong. 2021. CARM: Confidence-aware recommender model via review representation learning and historical rating behavior in the online platforms. Neurocomputing 455 (2021), 283–296.
Hongtao Liu, Yian Wang, Qiyao Peng, Fangzhao Wu, Lin Gan, Lin Pan, and Pengfei Jiao. 2020. Hybrid neural recommendation with joint deep representation learning of ratings and reviews. Neurocomputing 374 (2020), 77–85.
Junrui Liu, Tong Li, Mingliang Yu, Shiqiu Yang, Zifang Tang, and Zhen Yang. 2025. A Multi-Factor Collaborative Prediction for Review-based Recommendation. In Proceedings of the Nineteenth ACM RecSys. DOI: 10.1145/3705328.3748062
Jianmo Ni, Jiacheng Li, and Julian McAuley. 2019. Justifying Recommendations using Distantly-Labeled Reviews and Fine-Grained Aspects. In Proceedings o EMNLP-IJCNLP, Kentaro Inui, Jing Jiang, Vincent Ng, and Xiaojun Wan (Eds.). DOI: 10.18653/v1/D19-1018
Francesco Ricci, Lior Rokach, and Bracha Shapira. 2015. Recommender systems: introduction and challenges. Recommender systems handbook (2015), 1–34.
Md Shohel Sayeed, Varsha Roji, and Kalaiarasi Anbananthen. 2023. BERT: A Review of Applications in Sentiment Analysis. HighTech and Innovation Journal 4 (06 2023), 453–462. DOI: 10.28991/HIJ-2023-04-02-015
Mehdi Srifi, Ahmed Oussous, Ayoub Ait Lahcen, and Salma Mouline. 2020. Recommender systems based on collaborative filtering using review texts—a survey. Information 11, 6 (2020), 317.
NLP Town. 2019. bert-base-multilingual-uncased-sentiment. [link]
Heitor Werneck, Nícollas Silva, Matheus Carvalho Viana, Fernando Mourão, Adriano CM Pereira, and Leonardo Rocha. 2020. A survey on point-of-interest recommendation in location-based social networks. In Proceedings of the Brazilian Symposium on Multimedia and the Web. 185–192.
Frank Wilcoxon. 1992. Individual Comparisons by Ranking Methods. Springer New York, New York, NY, 196–202. DOI: 10.1007/978-1-4612-4380-9_16
Huiqiong Wu, Guibing Guo, Enneng Yang, Yudong Luo, Yadan Chu, Linying Jiang, and XingweiWang. 2024. PESI: Personalized Explanation recommendation with Sentiment Inconsistency between ratings and reviews. Knowledge-Based Systems 283 (2024), 111133.
Publicado
27/10/2026
Como Citar
VASCONCELOS, Naan; ANDRADE, Yan; PEREIRA, Adriano; ROCHA, Leonardo.
Quando a Nota Não Conta Toda a História: Inconsistências entre Ratings e Reviews em Sistemas de Recomendação. In: CONCURSO DE TRABALHOS DE INICIAÇÃO CIENTÍFICA - SIMPÓSIO BRASILEIRO DE SISTEMAS MULTIMÍDIA E WEB (WEBMEDIA), 32. , 2026, Lavras/MG.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2026
.
p. 66-69.
ISSN 2596-1683.
DOI: https://doi.org/10.5753/webmedia_estendido.2026.34371.
