A binary PSO-based ensemble under-sampling model for rebalancing imbalanced training data

Li, Jinyan; Wu, Yaoyang; Fong, Simon; Tallón‐Ballesteros, Antonio J.; Yang, Xin‐She; Mohammed, Sabah; Wu, Feng

doi:10.1007/s11227-021-04177-6

Cited by 25 publications

(5 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The undersampling method balances the dataset by removing some of the majority class samples [13]. Random undersampling is the most common type of undersampling algorithm, which randomly selects some majority class samples and combines them with the original minority class samples to form a balanced dataset [14].…”

Section: Undersamplingmentioning

confidence: 99%

HS-CGK: A Hybrid Sampling Method for Imbalance Data Based on Conditional Tabular Generative Adversarial Network and K-Nearest Neighbor Algorithm

Zhao,

Guan,

Xue

et al. 2024

cai

View full text Add to dashboard Cite

Class imbalance problem in datasets can lead to biased classification decisions in favor of majority class samples. Additionally, class overlap can cause fuzzy classification boundaries, affecting the performance of classification algorithms. To address these issues, we propose a hybrid sampling method based on conditional tabular generative adversarial network (CTGAN) and K-nearest neighbor (KNN) algorithm. Firstly, we introduce an oversampling algorithm, named DB-CTGAN, based on CTGAN. This algorithm filters noisy and boundary samples using the density-based spatial clustering of applications with noise (DBSCAN) clustering algorithm and generates synthetic samples that conform to the real data distribution using CTGAN. Finally, we combine the expanded fraudulent samples generated by DB-CTGAN with the normal samples and use the KNN overlap undersampling algorithm to remove the samples in the overlap region, solving the class overlap problem. Experimental results show that compared with eight sampling methods using four standard classification models (Random Forest, Decision Tree, Support Vector Classification, and XGBoost), the proposed method significantly improves the F1, AUC, and G-mean metrics on five real datasets.

show abstract

Section: Undersamplingmentioning

confidence: 99%

HS-CGK: A Hybrid Sampling Method for Imbalance Data Based on Conditional Tabular Generative Adversarial Network and K-Nearest Neighbor Algorithm

Zhao,

Guan,

Xue

et al. 2024

cai

View full text Add to dashboard Cite

show abstract

“…The purpose of undersampling is to ensure that the classifier is not biased toward the majority class and can learn patterns and make accurate predictions for both classes. Undersampling can help improve the model's ability to predict the minority class correctly [29].…”

Section: Datasetmentioning

confidence: 99%

A Novel Framework for Credit Card Fraud Detection

Mniai,

Tarik,

Jebari

2023

IEEE Access

View full text Add to dashboard Cite

Credit card transactions have grown considerably in the last few years. However, this increase has led to significant financial losses around the world. More than that, processing the enormous amount of generated data becomes very challenging, making the datasets highly dimensional and unbalanced. This means the collected data is suffering from two major problems. It is characterized by a severe difference in observation frequency between fraud and non-fraud transactions, and it contains irrelevant, inappropriate, and correlated data that negatively affects their prediction performance. Consequently, it has attracted the interest of machine learning (ML), which has become a significant actor in fraud detection. ML has provided methods such as Logistic Regression (LR), Support vector machines (SVM), Decision Trees (DT), Random Forest (RF), and K-Nearest Neighbors (KNN). However, these methods cannot meet the outstanding performance required to detect and predict unusual fraud patterns. In this regard, the contribution of this paper is to propose a framework for fraud detection (FFD). At first, to overcome the unbalanced data problem, the framework uses an undersampling technique. Next, a feature selection (FS) mechanism is applied to select only relevant features. Then, a Support Vector Data Description (SVDD) is used to build the ML model. SVDD aims to create a tight boundary around regular data points to distinguish them from potential outliers or anomalies. In order to enhance optimization capability for its hyperparameters C and σ, a modified version of the Particle Swarm Optimization (PSO) algorithm, Polynomial Self Learning PSO (PSLPSO), is proposed. As a result, the framework's effectiveness is shown in the experimental results on a real credit card transaction dataset.

show abstract

“…A novel ensemble method had been designed to have the benefit of using ensemble learning with the noble under-sampling method. The under-sampling was done through Binary PSO (BPSO) and it got the benefit of ensemble classifiers to find the right number of instances from the majority class 13 . The very most prevalent problem that the real-world dataset suffers from is class imbalance.…”

Section: Metaheuristic Algorithms For Instance Selectionmentioning

confidence: 99%

Handling the Class Imbalance Problem using an improved Sine Cosine Algorithm for Optimal Instance Selection

Moorthy

Arikumar²,

Prathiba

et al. 2023

Preprint

View full text Add to dashboard Cite

Class imbalance is a significant problem that is biased, exhibiting excellent performance towards the majority classes in the dataset while inhibiting inferior performance toward minority classes. When dealing with real-world issues, in particular healthcare problems, this kind of biased nature affects classification accuracy. Thus, a class imbalance is a danger that directly affects the effectiveness of any classification algorithm. The Improved Binary Sine Cosine Algorithm (IBSCA) has been used in this work to identify the subset of the majority class in the best possible way. The proposed IBSCA makes some enhancements over the conventional Binary Sine Cosine Algorithm (BSCA) to address the issue of premature convergence with the local optimal solutions. Intending to improve the classification accuracy for unbalanced datasets, the proposed IBSCA seeks to identify the optimal collection of instances from the majority class. The advised IBSCA uses a random agent’s location, which tends to devote considerable time to exploration to find the best possible set of instances. By using the geometric Mean (G-Mean) and F-Score to describe the fitness function, the proposed IBSCA aims to solve the multi-objective optimization issue. The most crucial metrics for assessing how well a classifier performs on skewed datasets are G-Meanand F-Score. Additionally, a V-shaped transfer function is used to handle the discrete nature of the class imbalance issue. On 18 datasets with different imbalance ratios taken from the KEEL repository, experimentation is conducted. Comparisons are made between the suggested IBSCA and the traditional BSCA, Binary Particle Swarm Optimization (BPSO), and Binary GreyWolf Optimization (BGWO). Additionally, the performance of the suggested IBSCA is evaluated against the top outcomes from different research papers. Metrics like Sensitivity, F-Score, G-Mean, and Area under Curve (AUC) show that the suggestedIBSCA outperforms the current algorithms. The statistical findings also demonstrate that the suggested IBSCA is more efficient than the other conventional algorithms.

show abstract

A binary PSO-based ensemble under-sampling model for rebalancing imbalanced training data

Cited by 25 publications

References 35 publications

HS-CGK: A Hybrid Sampling Method for Imbalance Data Based on Conditional Tabular Generative Adversarial Network and K-Nearest Neighbor Algorithm

HS-CGK: A Hybrid Sampling Method for Imbalance Data Based on Conditional Tabular Generative Adversarial Network and K-Nearest Neighbor Algorithm

A Novel Framework for Credit Card Fraud Detection

Handling the Class Imbalance Problem using an improved Sine Cosine Algorithm for Optimal Instance Selection

Contact Info

Product

Resources

About