Article
Computer Science, Information Systems
Chih-Fong Tsai, Wei-Chao Lin
Summary: The study treats class imbalance as anomaly detection problem, investigating the performance of OCC classifiers and their performance in ensemble learning. Results show that OCC classifiers perform well on datasets with high class imbalance ratios, but feature selection does not usually improve their performance, while combining multiple OCC classifiers can outperform individual classifiers.
Article
Health Care Sciences & Services
Vinod Kumar, Gotam Singh Lalotra, Ponnusamy Sasikala, Dharmendra Singh Rajput, Rajesh Kaluri, Kuruva Lakshmanna, Mohammad Shorfuzzaman, Abdulmajeed Alsufyani, Mueen Uddin
Summary: Healthcare is crucial for every individual, and clinical datasets play a significant role in developing intelligent healthcare systems. However, class imbalance in real-world datasets poses challenges in training classifiers. This study evaluates the performance of six classifiers on five imbalanced clinical datasets and explores different class balancing techniques. The results demonstrate the superiority of the SMOTEEN method among all the tested techniques.
Article
Biology
Manisha Saini, Seba Susan
Summary: Screening and diagnosis of diabetic retinopathy disease is a significant problem in the biomedical domain. The use of medical imagery from a patient's eye for computer-aided diagnosis has greatly advanced with the success of deep learning. However, challenges with imbalanced datasets, inconsistent annotations, limited samples, and inappropriate evaluation metrics have impacted the performance of deep learning models.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Computer Science, Artificial Intelligence
Bhagat Singh Raghuwanshi, Sanyam Shukla
Summary: This paper introduces a new variant of extreme learning machine, MCVCSELM, for effectively addressing binary class imbalance problems by utilizing minimum class variance and class-specific regularization. Experimental results demonstrate that the proposed algorithm outperforms several state-of-the-art methods for imbalanced learning.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Eyad Elyan, Carlos Francisco Moreno-Garcia, Chrisina Jayne
Summary: Class-imbalanced datasets are common in various domains, and using class decomposition and oversampling methods can effectively reduce the dominance of majority class instances. Experimental results demonstrate the effectiveness and superiority of the proposed hybrid approach in addressing class imbalance.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
William C. Sleeman, Bartosz Krawczyk
Summary: This paper proposes a compound framework for dealing with multi-class big data problems, addressing the existence of multiple classes and high volumes of data simultaneously. By analyzing instance-level difficulties in each class and embedding this information in popular resampling algorithms, informative balancing of multiple classes is achieved. Extensive experimental study shows that using instance-level information significantly improves learning from multi-class imbalanced big data.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Ashhadul Islam, Samir Brahim Belhaouari, Atiq Ur Rehman, Halima Bensmail
Summary: This study introduces an advanced algorithm called KNNOR to address class imbalance by studying the compactness and location of the minority class, identifying critical and safe areas for augmentation, and generating synthetic data points. Experimental results show that the proposed method consistently outperforms other state-of-the-art oversamplers on several common imbalanced datasets, making it easy to use and open source as a python library.
APPLIED SOFT COMPUTING
(2022)
Article
Computer Science, Artificial Intelligence
Cian Lin, Chih-Fong Tsai, Wei-Chao Lin
Summary: For imbalanced domain datasets, data re-sampling techniques such as under-sampling the majority class and over-sampling the minority class are often used to construct effective models. This study aims to determine the better order of combining under- and over-sampling methods. Experiment results show that if the under-sampling algorithm (IB3) is carefully chosen, further addition of the over-sampling step does not significantly improve performance. Moreover, performing instance selection first and over-sampling second with the IB3 algorithm yields the best results.
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Article
Biochemical Research Methods
Yuxuan Pang, Zhuo Wang, Jhih-Hua Jhong, Tzong-Yi Lee
Summary: The study investigated the physiochemical properties of anti-coronavirus peptides and established a classifier for identification. Imbalanced learning strategies were adopted to address the class-imbalance issue. A double-stages classifier was designed to identify anti-CoV peptides, showing promising results for assisting in the development of novel anti-CoV peptides.
BRIEFINGS IN BIOINFORMATICS
(2021)
Article
Chemistry, Multidisciplinary
Wei Liu, Jiaqing Mo, Furu Zhong
Summary: In this paper, a federated learning method combining regularization constraints and pseudo-label construction is proposed to address the issues of insufficient labeled data and model degradation in the case of data category imbalance. Experimental results show that the method improves the AUC by 7.35% and the average class sensitivity by 1.34% compared to state-of-the-art methods, indicating its strong learning capability on unbalanced datasets with small batches.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Artificial Intelligence
Wei Dai, Dayong Li, Ding Tang, Huamiao Wang, Yinghong Peng
Summary: Availability of large-scale annotated and class-balanced datasets is crucial for deep learning based computer vision tasks. This study proposes a framework for improving spot welding defects classification performance by using GAN-based data augmentation, including the generation of diverse minority-class images using balancing GAN and gradient penalty, and enhancing the classifier by adding the generated images to the training dataset.
Article
Computer Science, Artificial Intelligence
Tianlei Wang, Jiuwen Cao, Xiaoping Lai, Q. M. Jonathan Wu
Summary: Autoencoding is an important branch of representation learning in deep neural networks, and the newly developed WSI-AE and OCC algorithm based on it were experimentally proven to be effective in comparison with state-of-the-art AEs and OCC algorithms.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Mathematics
Dario Ramos-Lopez, Ana D. Maldonado
Summary: Multi-class classification in imbalanced datasets presents a challenging problem where traditional validation metrics may not be suitable. A cost-sensitive variable selection procedure is proposed to build a Bayesian network classifier, optimizing a specified cost function. Fine-tuning the objective validation function can improve prediction quality in imbalanced data or when considering asymmetric misclassification costs.
Article
Computer Science, Artificial Intelligence
Nuno Moniz, Vitor Cerqueira
Summary: In this paper, an Automated Imbalanced Classification method, ATOMIC, is proposed for imbalanced classification tasks. ATOMIC provides a ranking of solutions most likely to ensure an optimal approximation to a new domain, drastically reducing associated computational complexity and energy consumption by anticipating the loss of a large set of predictive solutions in new imbalanced learning tasks. Results demonstrate that the proposed method provides a relevant approach to imbalanced learning while reducing learning and testing efforts of candidate solutions by approximately 95%.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Multidisciplinary Sciences
Mingyang Deng, Yingshi Guo, Chang Wang, Fuwei Wu
Summary: The proposed oversampling method based on classification ranking and weight setting effectively addresses the oversampling problem of multi-class small samples, achieving balanced data distribution while maintaining the properties of the original samples. Compared to other algorithms, it also shows higher classification accuracy of around 90%, demonstrating practicality and generality for imbalanced multi-class samples.
Article
Computer Science, Artificial Intelligence
Hugo Jair Escalante, Manuel Montes-y-Gomez, Jesus A. Gonzalez, Pilar Gomez-Gil, Leopoldo Altamirano, Carlos A. Reyes, Carolina Reta, Alejandro Rosales
ARTIFICIAL INTELLIGENCE IN MEDICINE
(2012)
Article
Computer Science, Artificial Intelligence
Hugo Jair Escalante, Carlos A. Hernandez, Jesus A. Gonzalez, A. Lopez-Lopez, Manuel Montes, Eduardo F. Morales, L. Enrique Sucar, Luis Villasenor, Michael Grubinger
COMPUTER VISION AND IMAGE UNDERSTANDING
(2010)
Article
Astronomy & Astrophysics
R. Diaz-Hernandez, H. Peregrina-Barreto, L. Altamirano-Robles, J. A. Gonzalez-Bernal, A. E. Ortiz-Esquivel
EXPERIMENTAL ASTRONOMY
(2014)
Article
Computer Science, Artificial Intelligence
Luis Mena, Jesus A. Gonzalez, Gladys Maestre
Article
Computer Science, Artificial Intelligence
Alejandro Rosales-Perez, Jesus A. Gonzalez, Carlos A. Coello Coello, Hugo Jair Escalante, Carlos A. Reyes-Garcia
Proceedings Paper
Computer Science, Artificial Intelligence
Alejandro Rosales-Perez, Hugo Jair Escalante, Carlos A. Coello Coello, Jesus A. Gonzalez, Carlos A. Reyes-Garcia
2014 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC)
(2014)
Article
Computer Science, Artificial Intelligence
Jorge Morales, Jesus A. Gonzalez, Carlos A. Reyes-Garcia, Leopoldo Altamirano
INTELLIGENT DATA ANALYSIS
(2014)
Proceedings Paper
Computer Science, Artificial Intelligence
Alejandro Rosales-Perez, Carlos A. Reyes-Garcia, Pilar Gomez-Gil, Jesus A. Gonzalez, Leopoldo Altamirano
ADVANCES IN ARTIFICIAL INTELLIGENCE, PT I
(2011)