Denilson Alves Pereira
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Denilson Alves Pereira has a total of 55 co-authors. Here are the top co-authors by number of shared publications:
Berthier Ribeiro-Neto
Google Engineering, Belo Horizonte, Brazil
2 shared publicationsMarluce Rodrigues Pereira
Universidade Federal de Lavras (UFLA)
2 shared publications- Mozar José de Brito
Universidade Federal de Lavras, Brasil
2 shared publications Nivio Ziviani
Federal University of Minas Gerais, Belo Horizonte, Brazil
2 shared publicationsAhmed A. A. Esmin
Department of Computer Science, Federal University of Lavras, Brazil
1 shared publication- Ahmed A. A. Esmin
Department of Computer Science, Universidade Federal de Lavras, PO Box 3037, Lavras 37200-000, Brazil
1 shared publication
Latest Publications
- 30 September 2026
- Published by Springer Nature
- p. 170-185
On the Feasibility of Distilled Small Language Models for Agentic Text-to-SQL in Portuguese Under Real-World Conditions
- 30 September 2026
- Published by Springer Nature
- p. 170-185
On the Feasibility of Distilled Small Language Models for Agentic Text-to-SQL in Portuguese Under Real-World Conditions
- Open Access
- 8 September 2026
- Published by Sociedade Brasileira de Computação
- p. 71-84
DATAEXPLAINER: Comprehensible Agentic Data Science with Consumer-Grade Hardware
Data Science (DS), a complex field often involving advanced knowledge of statistics, mathematics, and programming, has historically been restricted to deeply knowledgeable human professionals. Recent advancements in artificial reasoning and the expanding toolset of large language models (LLMs) have challenged this paradigm. However, most LLM-based approaches for automated DS are still computationally demanding, time-consuming, and not especially easy to follow. In this work, we expand on previous advancements to present DATAEXPLAINER, a framework for automated DS that provides step-by-step solutions and surpasses the average Kaggle user while using a single consumer-grade GPU and less than 2 hours of processing.
- 8 September 2026
- Published by Sociedade Brasileira de Computação
- p. 71-84
Open AccessDATAEXPLAINER: Comprehensible Agentic Data Science with Consumer-Grade Hardware
Data Science (DS), a complex field often involving advanced knowledge of statistics, mathematics, and programming, has historically been restricted to deeply knowledgeable human professionals. Recent advancements in artificial reasoning and the expanding toolset of large language models (LLMs) have challenged this paradigm. However, most LLM-based approaches for automated DS are still computationally demanding, time-consuming, and not especially easy to follow. In this work, we expand on previous advancements to present DATAEXPLAINER, a framework for automated DS that provides step-by-step solutions and surpasses the average Kaggle user while using a single consumer-grade GPU and less than 2 hours of processing.
- Open Access
- 22 October 2025
- Published by Universidad Nacional de La Plata in Journal of Computer Science and Technology
- Vol. 25 (2)
Evaluation of Approaches Based on the BERT Model for Opinion Mining about the Cachaça Beverage
Opinion mining is a natural language processing task that aims to classify user opinions expressed on e-commerce platforms, social networks and other media. It is an important tool for decision making, monitoring products/services, detecting trends, developing marketing strategies, among others. Much research has been carried out addressing opinions in the English language. The Portuguese language is still very lacking in linguistic resources aimed at training machine learning models. This work contributes to the evaluation of approaches based on the BERT language model for opinion mining in the Portuguese language, in particular, by creating and evaluating a dataset with labeled data in the domain of the beverage called Cachaça. This is a popular drink in Brazil, and of great economic importance. As a result of the experimental evaluation, the approaches based on the BERT model stood out in relation to two baselines, and in a cross-domain evaluation, they achieved values greater than 0.97 in the F1 metric for classification into 2 classes and 0.64 for 3 classes, in the dataset labeled for the cachaça beverage.
- 22 October 2025
- Published by Universidad Nacional de La Plata in Journal of Computer Science and Technology
- Vol. 25 (2)
Open AccessEvaluation of Approaches Based on the BERT Model for Opinion Mining about the Cachaça Beverage
Opinion mining is a natural language processing task that aims to classify user opinions expressed on e-commerce platforms, social networks and other media. It is an important tool for decision making, monitoring products/services, detecting trends, developing marketing strategies, among others. Much research has been carried out addressing opinions in the English language. The Portuguese language is still very lacking in linguistic resources aimed at training machine learning models. This work contributes to the evaluation of approaches based on the BERT language model for opinion mining in the Portuguese language, in particular, by creating and evaluating a dataset with labeled data in the domain of the beverage called Cachaça. This is a popular drink in Brazil, and of great economic importance. As a result of the experimental evaluation, the approaches based on the BERT model stood out in relation to two baselines, and in a cross-domain evaluation, they achieved values greater than 0.97 in the F1 metric for classification into 2 classes and 0.64 for 3 classes, in the dataset labeled for the cachaça beverage.
- Open Access
- 19 May 2025
- Published by Sociedade Brasileira de Computação
- p. 260-269
Performance evaluation of LLMs in the Text-to-SQL task in Portuguese
Context: The rising need for consulting data in industry and academic contexts has fueled Text-to-SQL development, where natural language queries are translated into SQL, making data access easier. Problem: Most research focuses on English Text-to-SQL, leaving Portuguese—a language spoken by over 260 million people— underrepresented, creating challenges for organizations reliant on accurate data retrieval in Portuguese. Solution: This study evaluates the effectiveness of various Large Language Models (LLMs) on Portuguese Text-to-SQL tasks using a validated translation of the Spider benchmark. IS Theory: This research applies Task-Technology Fit (TTF) Theory to assess how well LLMs meet the needs of Portuguese Text-to-SQL tasks. TTF evaluates the match between LLM capabilities and task requirements, including language understanding, schema recognition, and SQL generation. Proper alignment is key for effective data retrieval, particularly for Portuguese-language applications in organizational decisionmaking. Method: A comparative analysis of seven LLMs—tested on both Portuguese and English Spider benchmarks—was performed. Exact Match (EM) and Execution Accuracy (EX) metrics measured performance, and a zero-shot prompting approach maintained consistency. Results Summary: Larger LLMs and specialized code models excelled, showing less performance variance between Portuguese and English tasks. Generalist models, however, produced verbose outputs, which may limit practical use in production systems. Contributions and Impact on IS: This research establishes baseline Portuguese Text-to-SQL metrics and insights into language adaptability in LLMs, offering guidance for organizations seeking Portuguese-language data solutions. By bridging language gaps, it advances data-driven practices and fosters growth in Portuguese-speaking regions.
- 19 May 2025
- Published by Sociedade Brasileira de Computação
- p. 260-269
Open AccessPerformance evaluation of LLMs in the Text-to-SQL task in Portuguese
Context: The rising need for consulting data in industry and academic contexts has fueled Text-to-SQL development, where natural language queries are translated into SQL, making data access easier. Problem: Most research focuses on English Text-to-SQL, leaving Portuguese—a language spoken by over 260 million people— underrepresented, creating challenges for organizations reliant on accurate data retrieval in Portuguese. Solution: This study evaluates the effectiveness of various Large Language Models (LLMs) on Portuguese Text-to-SQL tasks using a validated translation of the Spider benchmark. IS Theory: This research applies Task-Technology Fit (TTF) Theory to assess how well LLMs meet the needs of Portuguese Text-to-SQL tasks. TTF evaluates the match between LLM capabilities and task requirements, including language understanding, schema recognition, and SQL generation. Proper alignment is key for effective data retrieval, particularly for Portuguese-language applications in organizational decisionmaking. Method: A comparative analysis of seven LLMs—tested on both Portuguese and English Spider benchmarks—was performed. Exact Match (EM) and Execution Accuracy (EX) metrics measured performance, and a zero-shot prompting approach maintained consistency. Results Summary: Larger LLMs and specialized code models excelled, showing less performance variance between Portuguese and English tasks. Generalist models, however, produced verbose outputs, which may limit practical use in production systems. Contributions and Impact on IS: This research establishes baseline Portuguese Text-to-SQL metrics and insights into language adaptability in LLMs, offering guidance for organizations seeking Portuguese-language data solutions. By bridging language gaps, it advances data-driven practices and fosters growth in Portuguese-speaking regions.
- Open Access
- 19 May 2025
- Published by Sociedade Brasileira de Computação
- p. 449-458
Accessibility Assessment in Banking Chatbots using Texts Written by Deaf People who use Portuguese as a Second Language
Context: The study examines the ability of banking chatbots to understand and respond to interactions with deaf people who use Portuguese as a second language (L2), with a focus on digital accessibility in automated banking services. Problem: The hypothesis is that banking chatbots have difficulty interpreting requests in Portuguese as L2 and expressions based on the gloss of Brazilian Sign Language (Libras), resulting in digital exclusion and barriers in service. Solution: The proposal diagnoses the ability of chatbots, aiming at their understanding in the interactions of deaf users in order to improve accessibility. SI Theory: The research uses Inclusive Design and Task-Technology Fit (TTF), assessing whether chatbots are suitable for the needs of users who use Portuguese as L2. Method: A case study was carried out, with a qualitative approach. Ten banking activities were simulated in the chatbots of Banco do Brasil, Bradesco and Itaú, using inputs in formal Portuguese and as L2, with the help of ChatGPT for translations. Summary of Results: The results indicate error rates between 10% and 70%, with most errors in inputs that used Portuguese as L2 or colloquial language, evidencing the inadequacy of the chatbots for the deaf community. Contributions and Impact to the Field of IS: The study suggests the need for improvements in the language models of the chatbots to make them more inclusive for deaf users, generating impact in academia and in the technology industry.
- 19 May 2025
- Published by Sociedade Brasileira de Computação
- p. 449-458
Open AccessAccessibility Assessment in Banking Chatbots using Texts Written by Deaf People who use Portuguese as a Second Language
Context: The study examines the ability of banking chatbots to understand and respond to interactions with deaf people who use Portuguese as a second language (L2), with a focus on digital accessibility in automated banking services. Problem: The hypothesis is that banking chatbots have difficulty interpreting requests in Portuguese as L2 and expressions based on the gloss of Brazilian Sign Language (Libras), resulting in digital exclusion and barriers in service. Solution: The proposal diagnoses the ability of chatbots, aiming at their understanding in the interactions of deaf users in order to improve accessibility. SI Theory: The research uses Inclusive Design and Task-Technology Fit (TTF), assessing whether chatbots are suitable for the needs of users who use Portuguese as L2. Method: A case study was carried out, with a qualitative approach. Ten banking activities were simulated in the chatbots of Banco do Brasil, Bradesco and Itaú, using inputs in formal Portuguese and as L2, with the help of ChatGPT for translations. Summary of Results: The results indicate error rates between 10% and 70%, with most errors in inputs that used Portuguese as L2 or colloquial language, evidencing the inadequacy of the chatbots for the deaf community. Contributions and Impact to the Field of IS: The study suggests the need for improvements in the language models of the chatbots to make them more inclusive for deaf users, generating impact in academia and in the technology industry.
- 22 January 2025
- Published by Springer Nature in Language Resources and Evaluation
- Vol. 59 (3) , 2185-2211
UFLA-FORMS: an academic forms dataset for information extraction in the Portuguese language
- 22 January 2025
- Published by Springer Nature in Language Resources and Evaluation
- Vol. 59 (3) , 2185-2211
UFLA-FORMS: an academic forms dataset for information extraction in the Portuguese language
- 3 August 2024
- Published by Springer Nature in Neural Computing and Applications
- Vol. 36 (31) , 19353-19363
Approaches based on language models for aspect extraction for sentiment analysis in the Portuguese language
- 3 August 2024
- Published by Springer Nature in Neural Computing and Applications
- Vol. 36 (31) , 19353-19363
Approaches based on language models for aspect extraction for sentiment analysis in the Portuguese language
- 17 June 2023
- Published by Springer Nature in Language Resources and Evaluation
- Vol. 58 (4) , 1315-1333
CachacaNER: a dataset for named entity recognition in texts about the cachaça beverage
- 17 June 2023
- Published by Springer Nature in Language Resources and Evaluation
- Vol. 58 (4) , 1315-1333
CachacaNER: a dataset for named entity recognition in texts about the cachaça beverage
- 26 July 2020
- Published by Springer Nature in International Journal on Digital Libraries
- Vol. 21 (4) , 407-421
PVAF: an environment for disambiguation of scientific publication venues
- 26 July 2020
- Published by Springer Nature in International Journal on Digital Libraries
- Vol. 21 (4) , 407-421
PVAF: an environment for disambiguation of scientific publication venues
- 6 July 2020
- Published by Springer Nature in Artificial Intelligence Review
- Vol. 54 (2) , 1087-1115
A survey of sentiment analysis in the Portuguese language
- 6 July 2020
- Published by Springer Nature in Artificial Intelligence Review
- Vol. 54 (2) , 1087-1115
A survey of sentiment analysis in the Portuguese language