An Unsupervised Entropy-Complexity Framework for Real-Time Concept Drift Detection Keila Barbosa Costa, Alejandro C. Frery, André Aquino, George D. C. Cavalcanti IEEE Access, 2026 Monitoring and detecting changes in sensor and ubiquitous network data streams (data drifts) is essential to ensure reliability, adaptability, and security in dynamic environments. We present <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Quantile-based Change Voting with Entropy and Complexity</i> (QCV-HxC), an unsupervised framework for drift detection based on the temporal evolution of permutation entropy and statistical complexity, both derived from ordinal patterns representations. This framework combines smoothed temporal derivatives of these measures using composite functions, identifies change points using quantile-based thresholds, and then applies a voting mechanism across detection windows. We evaluate QCV-HxC on both real-world data with temporal dependencies (Electricity Market) and synthetic streams exhibiting abrupt concept shifts (RTG_2abrupt). Experimental results demonstrate that QCV-HxC achieves an average F<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub>-score of 0.80, surpassing traditional detectors such as ADWIN and Page-Hinkley-while operating entirely without label supervision. These findings indicate that the proposed framework provides a robust, efficient, and interpretable solution for real-time drift detection and online monitoring in wireless, sensor, and ubiquitous computing environments.
Univariate vs multivariate prediction for containerised applications auto-scaling: a comparative study Wellison R. M. Santos, Adalberto R. Sampaio, Nelson S. Rosa, George D. C. Cavalcanti Proceedings of the ACM Symposium on Applied Computing, 2025 Adaptive containerised systems have been developed using the Time Series Forecasting (TSF) technique. TSF analyses historical data patterns to estimate future trends, assuming they will occur again. Identifying future trends allows anticipating problems (e.g., high latency) and acting (e.g., replicating the service) to fix them before they occur. Depending on the number of features (i.e., metrics) used as input for prediction, TSF can be classified as univariate (single feature) or multivariate (two or more features). Despite the popularity of both TSF strategies, a unique strategy is typically implemented, and there is no comparison with the other. However, it is known that no strategy is the best choice for all possible scenarios. This paper presents a comparative study assessing univariate and multivariate proactive auto-scaling of containerised applications. A custom-made multivariate auto-scaling tool called Multivariate Forecasting Tool (MFT) was developed and compared with a production-grade univariate system called Predict Kube (PK). Both applications were evaluated using four popular open-source benchmark applications. The results show that the multivariate strategy decreased the response time of the evaluated applications in 75% of the experiments (i.e., 9 out of 12) compared to the univariate, and it was more cost-effective in half of them (i.e., 6 out of 12). Furthermore, they also indicate that the multivariate strategy efficiency is more significant as the number of containers composing the application increases. This comparative study is expected to be a helpful guide for developers who want to choose the most effective proactive approach for their auto-scaling solutions.
Multi-View Autoencoders for Fake News Detection Ingryd V. S. T. Pereira, George D. C. Cavalcanti, Rafael M. O. Cruz 2025 IEEE Symposium on Computational Intelligence in Natural Language Processing and Social Media Ci Nlpsome 2025, 2025 Given the volume and speed at which fake news spreads across social media, automatic fake news detection has become a highly important task. However, this task presents several challenges, including extracting textual features that contain relevant information about fake news. Research about fake news detection shows that no single feature extraction technique consistently outperforms the others across all scenarios. Nevertheless, different feature extraction techniques can provide complementary information about the textual data and enable a more comprehensive representation of the content. This paper proposes using multi-view autoencoders to generate a joint feature representation for fake news detection by integrating several feature extraction techniques commonly used in the literature. Experiments on fake news datasets show a significant improvement in classification performance compared to individual views (feature representations). We also observed that selecting a subset of the views instead of composing a latent space with all the views can be advantageous in terms of accuracy and computational effort. For further details, including source codes, figures, and datasets, please refer to the project's repository: https://github.com/ingrydpereira/multiview-fake-news.
PIPES: A Meta-dataset of Machine Learning Pipelines Cynthia Moreira Maia, Lucas B. V. de Amorim, George D. C. Cavalcanti, Rafael M. O. Cruz Proceedings of the International Joint Conference on Neural Networks, 2025 Solutions to the Algorithm Selection Problem (ASP) in machine learning face the challenge of high computational costs associated with evaluating various algorithms' performances on a given dataset. To mitigate this cost, the meta-learning field can leverage previously executed experiments shared in online repositories such as OpenML. OpenML provides an extensive collection of machine learning experiments. However, an analysis of OpenML’s records reveals limitations. It lacks diversity in pipelines, specifically when exploring data preprocessing steps/blocks, such as scaling or imputation, resulting in limited representation. Its experiments are often focused on a few popular techniques within each pipeline block, leading to an imbalanced sample. To overcome the observed limitations of OpenML, we propose PIPES, a collection of experiments involving multiple pipelines designed to represent all combinations of the selected sets of techniques, aiming at diversity and completeness. PIPES stores the results of experiments performed applying 9,408 pipelines to 300 datasets. It includes detailed information on the pipeline blocks, training and testing times, predictions, performances, and the eventual error messages. This comprehensive collection of results allows researchers to perform analyses across diverse and representative pipelines and datasets. PIPES also offers potential for expansion, as additional data and experiments can be incorporated to support the meta-learning community further. The data, code, supplementary material, and all experiments can be found at https://github.com/cynthiamaia/PIPES.git.
HSFN: Hierarchical Selection for Fake News Detection building Heterogeneous Ensemble Sara B. Coutinho, Rafael M. O. Cruz, Francimaria R. S. Nascimento, George D. C. Cavalcanti Conference Proceedings IEEE International Conference on Systems Man and Cybernetics, 2025 Psychological biases, such as confirmation bias, make individuals particularly vulnerable to believing and spreading fake news on social media, leading to significant consequences in domains such as public health and politics. Machine learning–based fact-checking systems have been widely studied to mitigate this problem. Among them, ensemble methods are particularly effective in combining multiple classifiers to improve robustness. However, their performance heavily depends on the diversity of the constituent classifiers—selecting genuinely diverse models remains a key challenge, especially when models tend to learn redundant patterns. In this work, we propose a novel automatic classifier selection approach that prioritizes diversity, also extended by performance. The method first computes pairwise diversity between classifiers and applies hierarchical clustering to organize them into groups at different levels of granularity. A HierarchySelect then explores these hierarchical levels to select one pool of classifiers per level, each representing a distinct intra-pool diversity. The most diverse pool is identified and selected for ensemble construction from these. The selection process incorporates an evaluation metric reflecting each classifier’s performance to ensure the ensemble also generalises well. We conduct experiments with 40 heterogeneous classifiers across six datasets from different application domains and with varying numbers of classes. Our method is compared against the Elbow heuristic and state-of-the-art baselines. Results show that our approach achieves the highest accuracy on two of six datasets. The implementation details are available on the project’s repository: https://github.com/SaraBCoutinho/HSFN.
DRES: Fake news detection by dynamic representation and ensemble selection Faramarz Farhangian, Leandro Augusto Ensina, George D C Cavalcanti, Rafael M. O. Cruz Emnlp 2025 2025 Conference on Empirical Methods in Natural Language Processing Proceedings of the Conference, 2025 The rapid spread of information via social media has made text-based fake news detection critically important due to its societal impact.This paper presents a novel detection method called Dynamic Representation and Ensemble Selection (DRES) for identifying fake news based solely on text.DRES leverages instance hardness measures to estimate the classification difficulty for each news article across multiple textual feature representations.By dynamically selecting the textual representation and the most competent ensemble of classifiers for each instance, DRES significantly enhances prediction accuracy.Extensive experiments show that DRES achieves notable improvements over state-of-the-art methods, confirming the effectiveness of representation selection based on instance hardness and dynamic ensemble selection in boosting performance.Codes and data are available at:
Music Genre Classification Using Contrastive Dissimilarity Gabriel Henrique Costanzi, Lucas O. Teixeira, Gustavo Z. Felipe, George D. C. Cavalcanti, Yandre M. G. Costa International Conference on Systems Signals and Image Processing, 2024
The dissimilarity approach: a review Yandre M. G. Costa, Diego Bertolini, Alceu S. Britto, George D. C. Cavalcanti, Luiz E. S. Oliveira Artificial Intelligence Review, 2020
Speaker segmentation using i-vector in meetings domain Leonardo V. Neri, Hector N.B. Pinheiro, Ing Ren Tsang, George D. da C. Cavalcanti, Andre G. Adami ICASSP IEEE International Conference on Acoustics Speech and Signal Processing Proceedings, 2017
Optimizing speaker-specific filter banks for speaker verification Hector N. B. Pinheiro, Fernando M. P. Neto, Adriano L. I. Oliveira, Tsang Ing Ren, George D. C. Cavalcanti, Andre G. Adami ICASSP IEEE International Conference on Acoustics Speech and Signal Processing Proceedings, 2017
Pixel Clustering for Face Recognition Tiago B.A. De Carvalho, Maria A.A. Sibaldo, Ing Ren Tsang, George D.C. Cavalcanti, Ing Jyh Tsang, Jan Sijbers Proceedings 2016 5th Brazilian Conference on Intelligent Systems Bracis 2016, 2017
An architecture to classify desertification areas using hyperspectral images and the optimum path forest algorithm Electronic Journal of Geotechnical Engineering, 2016
Supervised fractional eigenfaces T. B. A. de Carvalho, A. M. Costa, M. A. A. Sibaldo, I. R. Tsang, G. D. C. Cavalcanti Proceedings International Conference on Image Processing Icip, 2015
Fractional Eigenfaces T. B. A. de Carvalho, M. A. A. Sibaldo, I. R. Tsang, G. D. C. Cavalcanti, I. J. Tsang, J. Sijbers 2014 IEEE International Conference on Image Processing Icip 2014, 2014
A proposal for path loss prediction in urban environments using support vector regression Advanced International Conference on Telecommunications Aict, 2014
Choosing instance selection method using meta-learning Shayane de Oliveira Moura, Marcelo Bassani de Freitas, Halisson A. C. Cardoso, George D. C. Cavalcanti Conference Proceedings IEEE International Conference on Systems Man and Cybernetics, 2014
Pedestrian detection under progressive occlusion Silvio G.O. Santos, Tsang Ing Ren, George D.C. Cavalcanti, Tsang Ing Jyh, Jan Sijbers Proceedings 2013 IEEE International Conference on Systems Man and Cybernetics Smc 2013, 2013
Video colortoning Ing Ren Tsang, Diogo C. Lemos, Dario S.M. Pinheiro, George D.C. Cavalcanti, Ing Jyh Tsang Conference Proceedings IEEE International Conference on Systems Man and Cybernetics, 2012
A modular architecture based on image quality for fingerprint spoof detection George D. C. Cavalcanti, Luis Filipe A. Pereira, Hector N. B. Pinheiro, Jose Ivson S. Silva, Anderson G. Silva, Thais M. L. Pina, Daniel B. O. Carvalho, Tsang Ing Ren Conference Proceedings IEEE International Conference on Systems Man and Cybernetics, 2012
Application of the IPSONet in face detection Elliackin M. N. Figueiredo, Rafael G. Mesquita, Teresa B. Ludermir, George D. C. Cavalcanti Proceedings of the International Joint Conference on Neural Networks, 2012
A fingerprint spoof detection based on MLP and SVM Luis Filipe A. Pereira, Hector N. B. Pinheiro, Jose Ivson S. Silva, Anderson G. Silva, Thais M. L. Pina, George D. C. Cavalcanti, Tsang Ing Ren, Joao Paulo N. de Oliveira Proceedings of the International Joint Conference on Neural Networks, 2012
Iris segmentation and recognition using 2D log-Gabor filters Carlos A. C. M. Bastos, Tsang Ing Ren, George D. C. Cavalcanti Lecture Notes in Computer Science Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics, 2012
Real-time head pose estimation for mobile devices Euclides N. Arcoverde Neto, Rafael M. Barreto, Rafael M. Duarte, Joao Paulo Magalhaes, Carlos A. C. M. Bastos, Tsang Ing Ren, George D. C. Cavalcanti Lecture Notes in Computer Science Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics, 2012
BSOM network for pupil segmentation Gabriel S. Vasconcelos, Carlos A.C.M. Bastos, Tsang Ing Ren, George D.C. Cavalcanti Proceedings of the International Joint Conference on Neural Networks, 2011
Market volatility modeling for short time window Paulo S.G. de Mattos Neto, David A. Silva, Tiago A.E. Ferreira, George D.C. Cavalcanti Physica A Statistical Mechanics and Its Applications, 2011
Fuzzy active contour models Cesar Lima Pereira, Carlos A. C. M. Bastos, Tsang Ing Ren, George D. C. Cavalcanti IEEE International Conference on Fuzzy Systems, 2011
Mammographic images segmentation using texture descriptors A.A. Mascaro, C.A.B. Mello, W.P. Santos, G.D.C. Cavalcanti Proceedings of the 31st Annual International Conference of the IEEE Engineering in Medicine and Biology Society Engineering the Future of Biomedicine Embc 2009, 2009
An Unsupervised entropy-complexity Framework for Real-Time Concept Drift Detection K Barbosa, A Frery, A Aquino, GDC Cavalcanti IEEE Access , 2026 2026
Review of transformers applied to wind turbine data: Exploring the potential of transformers in revolutionizing wind turbine data analysis DAB Junior, GNP Leite, AAO Villa, ACA da Costa, O de Castro Vilela, ... Renewable and Sustainable Energy Reviews 232, 116801 , 2026 2026
MetaML: a multi-label meta-learning approach for pipeline recommendation CM Maia, LBV de Amorim, GDC Cavalcanti, RMO Cruz Machine Learning 114 (12), 278 , 2025 2025
Short-Term Wind Power Forecasting with Transformer-Based Models Enhanced by Time2Vec and Efficient Attention DA Bispo Junior, GNP Leite, EL Droguett, OVC de Souza, LA Lisboa, ... Energies 18 (23), 6162 , 2025 2025 Citations: 3
Perceptual Influence: Improving the Perceptual Loss Design for Low-Dose CT Enhancement GA Viana, LFA Pereira, TI Ren, GDC Cavalcanti, J Sijbers arXiv preprint arXiv:2509.23025 , 2025 2025
Detecção de Fake News em Português: Análise Comparativa entre Métodos de Representação em Português, Inglês e Multilíngues CB Vieira, JVS Souza, GDC Cavalcanti Brazilian Workshop on Social Network Analysis and Mining (BraSNAM), 187-199 , 2025 2025
Improving open set recognition with dissimilarity-based metric learning LO Teixeira, D Bertolini, LS Oliveira, GDC Cavalcanti, YMG Costa Knowledge-Based Systems, 114108 , 2025 2025 Citations: 5
PIPES: A meta-dataset of machine learning pipelines CM Maia, LBV de Amorim, GDC Cavalcanti, RMO Cruz 2025 International Joint Conference on Neural Networks (IJCNN), 1-8 , 2025 2025 Citations: 1
Contrastive Dissimilarity for Writer Identification: a proof of concept F Pignelli, LO Teixeira, D Bertolini, L Nanni, GDC Cavalcanti, YMG Costa 2025 32nd International Conference on Systems, Signals and Image Processing … , 2025 2025 Citations: 1
Triplet dissimilarity: a texture classification approach using dissimilarity and siamese networks LO Teixeira, D Bertolini, LS Oliveira, GDC Cavalcanti, YMG Costa Soft Computing 29 (11), 4725-4742 , 2025 2025 Citations: 5
Univariate vs multivariate prediction for containerised applications auto-scaling: a comparative study WR M. Santos, AR Sampaio Jr, NS Rosa, GDC Cavalcanti Proceedings of the 40th ACM/SIGAPP Symposium on Applied Computing, 1098-1105 , 2025 2025 Citations: 1
Imbalanced malware classification: an approach based on dynamic classifier selection JVS Souza, CB Vieira, GDC Cavalcanti, RMO Cruz 2025 IEEE Symposium on Computational Intelligence in Security, Defence and … , 2025 2025 Citations: 3
Multi-view autoencoders for fake news detection IVST Pereira, GDC Cavalcanti, RMO Cruz 2025 IEEE Symposium on Computational Intelligence in Natural Language … , 2025 2025 Citations: 2
Analysis of signals from air conditioner compressors with ordinal patterns and machine learning K Barbosa, AC Frery, GDC Cavalcanti Journal of Low Frequency Noise, Vibration and Active Control 44 (1), 21-38 , 2025 2025 Citations: 6
Gender bias detection on hate speech classification: an analysis at feature-level FRS Nascimento, GDC Cavalcanti, MD Costa-Abreu Neural Computing and Applications 37 (5), 3887-3905 , 2025 2025 Citations: 10
Short-Term Wind Power Forecasting with Transformer-Based Models Enhanced by Time2Vec and Efficient Attention DAB Junior, G de Novaes Pires Leite, EL Droguett, OVC de Souza, ... Energies 18 (23), 1-41 , 2025 2025
DRES: Fake news detection by dynamic representation and ensemble selection F Farhangian, LA Ensina, GDC Cavalcanti, RMO Cruz Conference on Empirical Methods in Natural Language Processing (EMNLP … , 2025 2025 Citations: 4
HSFN: Hierarchical Selection for Fake News Detection building Heterogeneous Ensemble SB Coutinho, RMO Cruz, FRS Nascimento, GDC Cavalcanti IEEE International Conference on Systems, Man, and Cybernetics (SMC) , 2025 2025
Triplet dissimilarity: improving dissimilarity approaches through metric learning LO Teixeira, D Bertolini, LES Oliveira, GDC Cavalcanti, YMG Costa Procedia Computer Science 264, 147-156 , 2025 2025
MOST CITED SCHOLAR PUBLICATIONS
The choice of scaling technique matters for classification performance LBV de Amorim, GDC Cavalcanti, RMO Cruz Applied Soft Computing 133, 109924 , 2023 2023 Citations: 560
Dynamic classifier selection: Recent advances and perspectives RMO Cruz, R Sabourin, GDC Cavalcanti Information Fusion 41, 195-216 , 2018 2018 Citations: 555
Assessing sentence scoring techniques for extractive text summarization R Ferreira, L de Souza Cabral, RD Lins, GP e Silva, F Freitas, ... Expert Systems with Applications 40 (14), 5755-5764 , 2013 2013 Citations: 396
META-DES: A Dynamic Ensemble Selection Framework using Meta-Learning RMO Cruz, R Sabourin, GDC Cavalcanti, TI Ren Pattern Recognition , 2015 2015 Citations: 330
Impact of lung segmentation on the diagnosis and explanation of COVID-19 in chest X-ray images LO Teixeira, RM Pereira, D Bertolini, LS Oliveira, L Nanni, ... Sensors 21 (21), 7116 , 2021 2021 Citations: 172
A graph-based friend recommendation system using genetic algorithm NB Silva, R Tsang, GDC Cavalcanti, J Tsang IEEE Congress on Evolutionary Computation (CEC), 233-239 , 2010 2010 Citations: 161
DESlib: A Dynamic ensemble selection library in Python RMO Cruz, LG Hafemann, R Sabourin, GDC Cavalcanti Journal of Machine Learning Research 21 (8), 1-5 , 2020 2020 Citations: 144
A study on combining dynamic selection and data preprocessing for imbalance learning A Roy, RMO Cruz, R Sabourin, GDC Cavalcanti Neurocomputing 286, 179-192 , 2018 2018 Citations: 139
Text line segmentation based on morphology and histogram projection RP dos Santos, GS Clemente, TI Ren, GDC Cavalcanti International Conference on Document Analysis and Recognition (ICDAR), 651-655 , 2009 2009 Citations: 135
Unsupervised Retinal Vessel Segmentation Using Combined Filters WS Oliveira, JV Teixeira, TI Ren, GDC Cavalcanti, J Sijbers PLOS ONE , 2016 2016 Citations: 133
Semi-supervised clustering for MR brain image segmentation NM Portela, GDC Cavalcanti, TI Ren Expert Systems with Applications 41 (4), 1492-1497 , 2014 2014 Citations: 118
META-DES.Oracle: Meta-learning and feature selection for dynamic ensemble selection RMO Cruz, R Sabourin, GDC Cavalcanti Information Fusion 38, 84-103 , 2017 2017 Citations: 115
Fake news detection: Taxonomy and comparative study F Farhangian, RMO Cruz, GDC Cavalcanti Information Fusion, 102140 , 2024 2024 Citations: 103
A Proposal for Path Loss Prediction in Urban Environments using Support Vector Regression R Timoteo, DC Cunha, GDC Cavalcanti AICT 2014, The Tenth Advanced International Conference on Telecommunications … , 2014 2014 Citations: 102
Online Pruning of Base Classifiers for Dynamic Ensemble Selection DVR Oliveira, GDC Cavalcanti, R Sabourin Pattern Recognition , 2017 2017 Citations: 92
A global-ranking local feature selection method for text categorization RHW Pinheiro, GDC Cavalcanti, RF Correa, TI Ren Expert Systems with Applications 39 (17), 12851-12857 , 2012 2012 Citations: 87
An approach to feature selection for keystroke dynamics systems based on PSO and feature weighting GLF Azevedo, GDC Cavalcanti, ECB Carvalho Filho IEEE Congress on Evolutionary Computation (CEC), 3577-3584 , 2007 2007 Citations: 84
Data-driven global-ranking local feature selection methods for text categorization RHW Pinheiro, GDC Cavalcanti, TI Ren Expert Systems with Applications 42 (4), 1941–1949 , 2015 2015 Citations: 74
FIRE-DES++: Enhanced Online Pruning of Base Classifiers for Dynamic Ensemble Selection RMO Cruz, DVR Oliveira, GDC Cavalcanti, R Sabourin Pattern Recognition 85, 149-160 , 2019 2019 Citations: 70
