Minority Class Samples Research Articles

Elevator Fault Diagnosis Method Based on IAO-XGBoost under Unbalanced Samples

Elevators are essential tools in daily life; timely and accurate fault diagnosis plays a crucial role in ensuring their safe operation. However, the existing elevator fault diagnosis methods often neglect the imbalance between the actual collected normal samples and the fault samples, resulting in low diagnostic accuracy. In this study, we propose an improved Aquila optimizer (IAO) extreme gradient boosting tree (XGBoost)-based elevator fault diagnosis method under unbalanced samples. The proposed method includes three main components: multi-domain feature extraction, sample balancing, and fault diagnosis. In the feature extraction phase, the time domain, frequency domain and entropy features of the vibration signal are extracted. In the sample balance phase, aiming at the problem of unbalanced fault samples, after feature selection using recursive feature elimination (RFE), the minority class samples are oversampled by applying SMOTE-Tomek. In the fault diagnosis phase, IAO is used to optimize the hyperparameters in the XGBoost, and the optimized hyperparameters are brought into XGBoost for fault diagnosis. The fault diagnosis accuracy of the method proposed in this study can reach 99.06%, and the method can accurately identify the fault state of the elevator.

Three-Stage Sampling Algorithm for Highly Imbalanced Multi-Classification Time Series Datasets

To alleviate the data imbalance problem caused by subjective and objective factors, scholars have developed different data-preprocessing algorithms, among which undersampling algorithms are widely used because of their fast and efficient performance. However, when the number of samples of some categories in a multi-classification dataset is too small to be processed via sampling or the number of minority class samples is only one or two, the traditional undersampling algorithms will be less effective. In this study, we select nine multi-classification time series datasets with extremely few samples as research objects, fully consider the characteristics of time series data, and use a three-stage algorithm to alleviate the data imbalance problem. In stage one, random oversampling with disturbance items is used to increase the number of sample points; in stage two, on the basis of the latter operation, SMOTE (synthetic minority oversampling technique) oversampling is employed; in stage three, the dynamic time-warping distance is used to calculate the distance between sample points, identify the sample points of Tomek links at the boundary, and clean up the boundary noise. This study proposes a new sampling algorithm. In the nine multi-classification time series datasets with extremely few samples, the new sampling algorithm is compared with four classic undersampling algorithms, namely, ENN (edited nearest neighbours), NCR (neighborhood cleaning rule), OSS (one-side selection), and RENN (repeated edited nearest neighbors), based on the macro accuracy, recall rate, and F1-score evaluation indicators. The results are as follows: of the nine datasets selected, for the dataset with the most categories and the fewest minority class samples, FiftyWords, the accuracy of the new sampling algorithm was 0.7156, far beyond that of ENN, RENN, OSS, and NCR; its recall rate was also better than that of the four undersampling algorithms used for comparison, corresponding to 0.7261; and its F1-score was 200.71%, 188.74%, 155.29%, and 85.61% better than that of ENN, RENN, OSS, and NCR, respectively. For the other eight datasets, this new sampling algorithm also showed good indicator scores. The new algorithm proposed in this study can effectively alleviate the data imbalance problem of multi-classification time series datasets with many categories and few minority class samples and, at the same time, clean up the boundary noise data between classes.

Production prediction modeling of food waste anaerobic digestion for resources saving based on SMOTE-LSTM

The global energy shortage and resource waste are becoming more and more prominent. With the massive production of food waste, anaerobic digestion through food waste is a key way to solve the resource shortage problem. To better study the anaerobic digestion process of food waste, a novel production prediction model of food waste process based on a long short-term memory (LSTM) method integrating the synthetic minority oversampling technique (SMOTE) based data expansion method is proposed. The minority class samples are analyzed and extended using the SMOTE, which are used as inputs of the LSTM. Then, the production prediction model can be built to reduce the influence of a few samples on the prediction model. Finally, the proposed method is applied in the methane production prediction model of actual food waste process plants. Compared with the back Propagation (BP) neural network, the extreme learning machine (ELM), the radial basis function (RBF) neural network, the support vector machine (SVM), the LSTM and the convolutional neural network (CNN), the experimental results have verified the higher applicability of the proposed method for the methane prediction result including an accuracy of 99.75% and the highest R2 of 0.9913 with minimal training and generalization errors. Moreover, by analyzing the prediction result and the actual methane production, the proposed method can effectively guide and timely adjustment the feed allocation for increasing the methane production per m3 of feed by 25.77%.

Improving Software Defect Prediction in Noisy Imbalanced Datasets

Software defect prediction is a popular method for optimizing software testing and improving software quality and reliability. However, software defect datasets usually have quality problems, such as class imbalance and data noise. Oversampling by generating the minority class samples is one of the most well-known methods to improving the quality of datasets; however, it often introduces overfitting noise to datasets. To better improve the quality of these datasets, this paper proposes a method called US-PONR, which uses undersampling to remove duplicate samples from version iterations and then uses oversampling through propensity score matching to reduce class imbalance and noise samples in datasets. The effectiveness of this method was validated in a software prediction experiment that involved 24 versions of software data in 11 projects from PROMISE in noisy environments that varied from 0% to 30% noise level. The experiments showed a significant improvement in the quality of datasets pre-processed by US-PONR in noisy imbalanced datasets, especially the noisiest ones, compared with 12 other advanced dataset processing methods. The experiments also demonstrated that the US-PONR method can effectively identify the label noise samples and remove them.

An Attack Detection Method for Self-Powered Sensor IoTs Based on Ensemble Learning

Aiming at the problem that machine learning has weak generalization ability and underfitting of minority class samples in network attack identification for self-powered sensor IoTs, this paper proposes an attack identification method based on ensemble learning. In the proposed method, a two-layer ensemble learning structure consisting of a binary random forest and the quintuple random forest is designed to improve the recognition accuracy; a binary grid search parameter tuning method is proposed to optimize the hyperparameters of the model to deal with the issue of the weak generalization ability of attack recognition; through random attributes combination and KNN algorithm, samples of the minor class are generated for unbalanced processing to address the underfitting problem of minority class identification; Gini impurity is adopted as the basis for feature selection, which reduces the complexity of model training and improves the efficiency of identification. The experimental results show that the network attack identification method for self-powered sensor IoTs based on ensemble learning proposed in this paper achieves 99.98% in terms of accuracy, precision, recall, and F1 measure.

Highly imbalanced fault classification of wind turbines using data resampling and hybrid ensemble method approach

Deep learning-based incipient fault diagnostic techniques have achieved surprisingly well in wind turbines. Due to component failures, wind turbines must undergo active maintenance, substantially influencing revenue and power generation. Unfortunately, there are consistently uneven data distributions between samples with faults and those without faults, resulting in incorrect fault classification. Wind turbine fault classification has a significant data imbalance problem, compromising learning attention for majority and minority classes. Machine learning methodologies based on Generative Adversarial Networks (GAN), over-sampling, and under-sampling techniques for generating synthetic data have been widely employed to address the imbalance data problem. However, the traditional synthetic minority oversampling technique (SMOTE) accomplishes oversampling using linear interpolation between close minority class samples, which could be confusing, subpar, and indistinguishable from the majority class. This study suggests combining over and under-sampling using adaptive SMOTE and edited nearest neighbors (ASMOTE-ENN) that incorporate over-sampling with adaptive SMOTE and under-sampling with ENN to improve the quality of the generated samples. With this resampling technique, noise in an imbalanced dataset is reduced on three levels by using an adaptive nearest neighbor selection algorithm to find the nearest neighbors that are visible. Then use SMOTE to create samples that precisely fall into the minority class, and later use the ENN technique to eliminate instances that contribute to noise afterwards. Resampling data created by combining over- and under-sampling approaches to match the data distribution over all classes is the foundation of the suggested method’s efficacy. A hybrid ensemble method is used for effective classification, including boosting, bagging, and stacking techniques. The original unbalanced and balanced data using the ASMOTE-ENN algorithm were classified using the proposed hybrid ensemble method. The classification results show that the proposed strategy is more accurate than a few imbalanced fault diagnosis techniques.

An imbalanced classification approach for establishment of cause-effect relationship between Heart-Failure and Pulmonary Embolism using Deep Reinforcement Learning

Pulmonary Embolism (PE) — a non-cardiac cause of cardiac arrest is a strenuous job to perform as it is non-specific in presentation and shares various overlapping features with various other clinical disorders like Myocardial Infarction, Pneumonia, etc. This paper delineates a method to identify Pulmonary Embolism (PE) as a reason for Heart Failure using Deep Reinforcement Learning (DRL) algorithm. The methodology has been partitioned into two phases. Phase I acts towards the creation of a novel dataset as there was no data available for this particular problem statement. Phase II deals with the scope and application of the DRL algorithm on the above-invented dataset. The dataset formed is imbalanced in its essence. To effectively tackle the imbalanced dataset challenge, a cutting-edge imbalanced classification algorithm rooted in the power of Deep Reinforcement Learning (DRL) has been embraced. The classification model has been drawn up as a Sequential Decision-Making procedure and is resolved by making use of the DRL algorithm viz Double Deep Q-Network (DDQN). A classification action is performed by an agent on a single sample at each time step. Classification actions are assessed by the environment, based on the assessment, a reward is given to an agent. In order to make the agent more responsive towards the minority class samples, rewards belonging to the minority class are higher. Under the supervision of this reward system, the agent finally learns the optimal classification policy for this imbalanced dataset problem. Comparative analysis of the DDQN algorithm with other Deep Reinforcement Learning algorithms, including Dueling DQN, and Proximal Policy Optimization (PPO) algorithms has been performed in order to arrive at the best learning technique for the dataset. Furthermore, the DDQN algorithm has been evaluated against a selection of Machine Learning (ML) algorithms. From the experimentation, it is demonstrated that the DDQN model outperformed other approaches with an accuracy of 99.99%, G-mean 0.9999, Recall 1.0000, and Specificity 0.9999.

Improving prediction model robustness with virtual sample construction for near-infrared spectra analysis

In a qualitative analysis of near-infrared spectroscopy (NIRS), when the samples to be analyzed are difficult to obtain or there are few counterexamples, the robustness of the models is poor, resulting in the decline of the generalization ability of the models. In this case, the effective method is to construct virtual samples to achieve the balance of categories. In this contribution, three virtual spectrum construction strategies including Synthetic Minority Oversampling Technique (SMOTE), Adaptive Synthetic Sampling (ADASYN), and Deep Convolutional Generative Adversarial Network (DCGAN) were explored to deal with the problem of insufficient or imbalanced sample numbers in NIRS analysis. The strategies were tested with the melamine and Yali pears two spectral datasets. The PLS-DA and Correct Recognition Rate (CRR) were used for discriminant model construction and accuracy evaluation, respectively. The results show that SMOTE, ADASYN, and DCGAN processing strategies can all improve the global CRR (CRRglob). The SMOTE and ADASYN can improve the CRR for majority class sample (CRRmaj), but the CRR for minority class sample (CRRmin) has decreased. For the DCGAN method, the CRRglob, CRRmaj, and CRRmin were all improved. The standard deviation of the results of the multiple parallel calculations demonstrates the robustness of DCGAN generation method. Therefore, the DCGAN method has good reliability and practicability, and can increase the robustness and generalization ability of the NIRS model.

FedVAE: Communication-Efficient Federated Learning With Non-IID Private Data

Federated learning (FL), collaboratively training a shared global model without exchanging and centralizing local data, provides a promising solution for privacy preserving. On the other hand, it is faced with two main challenges: First, high communication cost, and second, low model quality due to imbalanced or nonindependent and identically distributed (non-IID) data. In this article, we propose FedVAE, an FL framework based on variational autoencoder (VAE) for remote patient monitoring. FedVAE contains two lightweight VAEs: one for projecting data onto a lower dimensional space with similar distribution so as to alleviate the issues of excessive communication overhead and slow convergence rate caused by non-IID data, and the other for shunning training bias due to imbalanced data distribution through generating minority class samples. In general, the proposed FedVAE can improve the overall performance of FL models while consuming only a small amount of communication bandwidth. The experimental results show that the area under the curve (AUC) value of FedVAE can reach 0.9937, which is even higher than that of the traditional centralized model (0.9931). Besides, fine-tuning the global model with personalization can raise the average AUC to 0.9947. Moreover, compared with vanilla FL, FedVAE shows 0.87% improvement in AUC while reducing communication traffic by at least 95%.

Crash injury severity prediction considering data imbalance: A Wasserstein generative adversarial network with gradient penalty approach

For each road crash event, it is necessary to predict its injury severity. However, predicting crash injury severity with the imbalanced data frequently results in ineffective classifier. Due to the rarity of severe injuries in road traffic crashes, the crash data is extremely imbalanced among injury severity classes, making it challenging to the training of prediction models. To achieve interclass balance, it is possible to generate certain minority class samples using data augmentation techniques. Aiming to address the imbalance issue of crash injury severity data, this study applies a novel deep learning method, the Wasserstein generative adversarial network with gradient penalty (WGAN-GP), to investigate a massive amount of crash data, which can generate synthetic injury severity data linked to traffic crashes to rebalance the dataset. To evaluate the effectiveness of the WGAN-GP model, we systematically compare performances of various commonly-used sampling techniques (random under-sampling, random over-sampling, synthetic minority over-sampling technique and adaptive synthetic sampling) with respect to dataset balance and crash injury severity prediction. After rebalancing the dataset, this study categorizes the crash injury severity using logistic regression, multilayer perceptron, random forest, AdaBoost and XGBoost. The AUC, specificity and sensitivity are employed as evaluation indicators to compare the prediction performances. Results demonstrate that sampling techniques can considerably improve the prediction performance of minority classes in an imbalanced dataset, and the combination of XGBoost and WGAN-GP performs best with an AUC of 0.794 and a sensitivity of 0.698. Finally, the interpretability of the model is improved by the explainable machine learning technique SHAP (SHapley Additive exPlanation), allowing for a deeper understanding of the effects of each variable on crash injury severity. Findings of this study shed light on the prediction of crash injury severity with data imbalance using data-driven approaches.

A fault diagnosis method based on hybrid sampling algorithm with energy entropy under unbalanced conditions

For the degraded performance of the fault diagnosis model caused by massive normal samples and scarce fault samples under unbalanced conditions, a new fault diagnosis method based on a hybrid sampling algorithm and energy entropy, namely HSEEFD is proposed in this paper. In the proposed method, Empirical Modal Decomposition is employed to decompose the vibration signals into Intrinsic Mode Functions (IMFs), and the energy entropy feature of each IMF component is extracted to construct a feature vector matrix. Then, a new hybrid sampling algorithm using Tomek’s Links algorithm, Euclidean distance, K-means algorithm, and synthetic minority over-sampling technique (SMOTE), namely TSHSA is designed to balance the extracted features. Tomek’s Links algorithm is used to identify and remove the confusable majority class samples at the boundary. Euclidean distance is applied to find the suspected noise points in minority class samples and remove them. The k-means algorithm is employed to cluster the minority class samples and SMOTE is used to deal with each cluster according to the density of the clusters to synthesize new features. Finally, the support vector machine is applied to classify faults and realize fault diagnosis. The experiment results on the actual imbalanced data show that the proposed HSEEFD method can effectively improve the accuracy (AUC) of the fault diagnosis under unbalanced conditions by increasing the AUC value by more than 2.1%, and the AUC and G-mean by more than 0.7%, 2.1%, respectively.

A Data Enhancement Algorithm for DDoS Attacks Using IoT.

With the rapid development of the Internet of Things (IoT), the frequency of attackers using botnets to control IoT devices in order to perform distributed denial-of-service attacks (DDoS) and other cyber attacks on the internet has significantly increased. In the actual attack process, the small percentage of attack packets in IoT leads to low accuracy of intrusion detection. Based on this problem, the paper proposes an oversampling algorithm, KG-SMOTE, based on Gaussian distribution and K-means clustering, which inserts synthetic samples through Gaussian probability distribution, extends the clustering nodes in minority class samples in the same proportion, increases the density of minority class samples, and improves the amount of minority class sample data in order to provide data support for IoT-based DDoS attack detection. Experiments show that the balanced dataset generated by this method effectively improves the intrusion detection accuracy in each category and effectively solves the data imbalance problem.

Imbalanced least squares regression with adaptive weight learning

Least squares regression (LSR) has demonstrated promising performance in various classification tasks owing to its effectiveness and efficiency. However, there are some deficiencies that seriously hinder its application in imbalanced data scenarios. The first is that LSR strongly relies on a balanced class distribution. A severely imbalanced class distribution may seriously damage the effectiveness of the algorithm. Second, the utilized binary label matrix in the conventional LSR model may be too strict to learn a discriminative transformation matrix for imbalanced learning. To address the above issues, in this paper, an adaptive weight learning mechanism and label relaxation constraint are proposed and incorporated into the framework of LSR to tackle the imbalanced classification problem. The weight of each sample can be adaptively obtained according to the original distribution information of the imbalanced data, in which the importance of minority class samples can be better reflected with larger weights. A new label relaxation matrix consisting of the original label matrix and auxiliary matrix is constructed to widen the margins between different classes. Further, we provide an iterative algorithm with fast convergence to solve the resulting optimization problem. Extensive experimental results on diverse binary-class and multi-class imbalanced datasets show that the proposed method outperforms many other state-of-the-art imbalanced learning approaches.

Fault Diagnosis of Rolling Bearings in Primary Mine Fans under Sample Imbalance Conditions.

Rolling bearings are crucial parts of primary mine fans. In order to guarantee the safety of coal mine production, primary mine fans commonly work during regular operation and are immediately shut down for repair in case of failure. This causes the sample imbalance phenomenon in fault diagnosis (FD), i.e., there are many more normal state samples than faulty ones, seriously affecting the precision of FD. Therefore, the current study presents an FD approach for the rolling bearings of primary mine fans under sample imbalance conditions via symmetrized dot pattern (SDP) images, denoising diffusion probabilistic models (DDPMs), the image generation method, and a convolutional neural network (CNN). First, the 1D bearing vibration signal was transformed into an SDP image with significant characteristics, and the DDPM was employed to create a generated image with similar feature distributions to the real fault image of the minority class. Then, the generated images were supplemented into the imbalanced dataset for data augmentation to balance the minority class samples with the majority ones. Finally, a CNN was utilized as a fault diagnosis model to identify and detect the rolling bearings' operating conditions. In order to assess the efficiency of the presented method, experiments were performed using the regular rolling bearing dataset and primary mine fan rolling bearing data under actual operating situations. The experimental results indicate that the presented method can more efficiently fit the real image samples' feature distribution and generate image samples with higher similarity than other commonly used methods. Moreover, the diagnostic precision of the FD model can be effectively enhanced by gradually expanding and enhancing the unbalanced dataset.

WCGAN-GP based synthetic attack data generation with GA based feature selection for IDS

Cyber-attack is one of the alarming issues in today's era. Firewalls, Intrusion Detection Systems (IDSs), and other techniques are popularly applied to prevent those attacks. Intrusion Detection System (IDS) works as the watchdog to monitor the network traffic and identify an alert when an unauthorized action is detected. However, most IDSs are designed on traditional datasets, which do not address the data imbalance problem. As a result, their performance is not good enough against minority class samples. In order to overcome the aforementioned problem and detect cyber-attacks with higher precision, this paper proposes a novel IDS applying Wasserstein Conditional Generative Adversarial Network – Gradient Penalty (WCGAN-GP) and Genetic Algorithm (GA). The objective of the two components- WCGAN-GP and Genetic Algorithm are introduced in the proposed work for synthetic data generation, feature optimization, and attack detection system. The proposed model generates synthetic samples that follow the underlying data distribution of actual data. This paper uses a novel fitness function for the convergence of the Genetic algorithm and produces an optimal feature vector for the classification problem. In this work, an extensive experimental study applying NSL-KDD and UNSW-NB15 datasets in combination with generated samples and reduced feature set from the proposed model is carried out to evaluate the performance of different machine learning models. Observations show that Xgboost gives the best performances among others. To justify using a genetic algorithm for feature optimization, this paper has compared the proposed model with other traditional feature selection approaches such as PCA, Autoencoder, and T-SNE on the proposed model. It is found that the proposed GA-based feature selection outperforms others. The reason behind it is that the GA-based feature selection technique offers certain advantages over PCA and Autoencoder-based dimensionality reduction techniques, such as the ability to handle non-linearity between features in the data, enhanced flexibility in feature selection, interpretability, and human insight, handling high-dimensional data, robustness to noise and outliers. It also offers an advantage over T-SNE, a nonlinear dimensionality reduction technique in terms of computational efficiency. Furthermore, the model is also compared with other state-of-the-art which address data imbalance problems. Our proposed model with Xgboost classifier gives better results than other existing approaches in terms of accuracy (95.54%), precision (92.61%), recall (95.54%), F1-score (93.41%), and false alarm rate (4.30%) on the NSL-KDD dataset and accuracy (89.58%), precision (89.46%), recall (89.58%), F1-score (88.89%) and false alarm rate (1.18%) on the UNSW-NB15 dataset.

SPE-ACGAN: A Resampling Approach for Class Imbalance Problem in Network Intrusion Detection Systems

Network Intrusion Detection Systems (NIDSs) play a vital role in detecting and stopping network attacks. However, the prevalent imbalance of training samples in network traffic interferes with NIDS detection performance. This paper proposes a resampling method based on Self-Paced Ensemble and Auxiliary Classifier Generative Adversarial Networks (SPE-ACGAN) to address the imbalance problem of sample classes. To deal with the class imbalance problem, SPE-ACGAN oversamples the minority class samples by ACGAN and undersamples the majority class samples by SPE. In addition, we merged the CICIDS-2017 dataset and the CICIDS-2018 dataset into a more imbalanced dataset named CICIDS-17-18 and validated the effectiveness of the proposed method using the three datasets mentioned above. SPE-ACGAN is more effective than other resampling methods in improving NIDS detection performance. In particular, SPE-ACGAN improved the F1-score of Random Forest, CNN, GoogLeNet, and CNN + WDLSTM by 5.59%, 3.75%, 3.60%, and 3.56% after resampling.

RETRACTED ARTICLE: Fusion of transformer and ML-CNN-BiLSTM for network intrusion detection

Network intrusion detection system (NIDS) can effectively sense network attacks, which is of great significance for maintaining the security of cyberspace. To meet the requirements of efficient and accurate network status monitoring, a NIDS model using Transformer-based fusion deep learning architecture is proposed. Firstly, GAN-Cross is used to expand minority class sample data, thereby alleviating the issues of imbalanced minority class about the original dataset. Then, the Transformer module is used to adjust the ML-CNN-BiLSTM model to enhance the feature encoding ability of the intrusion model. Finally, the data enhancement model and feature enhancement model are integrated into the NIDS model, the detection model is optimized, the features of network state data are extracted at a deeper level, and the generalization ability of the detection model is enhanced. Some simulation experiments using UNSW-NB15 datasets show that the proposed fusion architecture can achieve efficient analysis of complex network traffic datasets, with an accuracy of 0.903, effectively improving the detection accuracy of NIDS and its ability to detect unknown attacks. The proposed model has good application value in ensuring the stable operation of network systems.

Identification of Customer Churn Considering Difficult Case Mining

In the process of user churn modeling, due to the imbalance between lost users and retained users, the use of traditional classification models often cannot accurately and comprehensively identify users with churn tendency. To address this issue, it is not sufficient to simply increase the misclassification cost of minority class samples in cost-sensitive methods. This paper proposes using the Focal Loss hard example mining technique to add the class weight α and the focus parameter γ to the cross-entropy loss function of LightGBM. In addition, it emphasizes the identification of customers at risk of churning and raises the cost of misclassification for minority and difficult-to-classify samples. On the basis of the preceding ideas, the FocalLoss_LightGBM model is proposed, along with random forests, SVM, XGBoost, and LightGBM. Empirical analysis based on a dataset of credit card users publicly available on the Kaggle website. The AUC, TPR, and G-mean index values were superior to the existing model, which can effectively improve the accuracy and stability of potential lost users.

An empirical hybridized Siamese network using hypercube natural aggregation algorithm for handling imbalance data learning

AbstractDealing with imbalanced data is a common challenge in machine learning, where one class has significantly fewer examples than another. Successfully addressing this challenge requires careful consideration of the data, algorithm, and evaluation metrics to ensure that the model accurately predicts the minority class. In this study, we present a hybrid approach called Siamese‐HYNAA, which combines a Siamese network and a population‐based optimizer hypercube natural aggregation algorithm (HYNAA) to generate candidate solutions for augmenting the minority class. We collected 10 imbalanced datasets ranging from 1.81 to 8.78 imbalanced ratios and built solution pairs based on correctly predicted candidate solutions using support vector machine (SVM). We then fed these solutions to the Siamese network, which employs a one‐shot learning approach to improve predictions with fewer candidate solutions. However, we found that SVM predicted only a small number of minority class samples accurately, prompting us to optimize the number of candidate solution pairs using HYNAA to generate more synthetic samples for the Siamese network. We evaluated our proposed strategy against basic SMOTE and our previous work, SMOTE‐PSOEV, using various performance measures, including ROC‐AUC learning curves, sensitivity, specificity, accuracy, Characteristic stability index, balanced accuracy, F1‐score, informedness, markedness, and execution time. Our results indicate that Siamese‐HYNAA generates promising results for imbalanced data.

Fault Diagnosis Method Based on CND-SMOTE and BA-SVM Algorithm

The problem of unbalanced data classification has gotten extensive attention in the past few years. Unbalanced sample data makes the fault diagnosis and classification accuracy rate low, and the capability to classify minority-class fault samples is restricted. To address the problem that the classification algorithm in machine learning has the insufficient capability to identify minority class samples for unbalanced sample data classification problems. Therefore, this paper proposes an improved support vector machine (SVM) classification method based on the synthetic minority over-sampling technique (SMOTE). For the sampler, an improved synthetic minority over-sampling technique based on the characteristics of neighborhood distribution (CND-SMOTE) algorithm is used to equilibrate the minority class samples and the majority class samples. For the classifier, the parameter optimization method of support vector machines based on the bat algorithm (BA-SVM) is used to solve the multi-classification problem of faulty samples. Finally, experimental results prove that the CND-SMOTE+BA-SVM algorithm can synthesize high-quality minority fault samples, increase the classification accuracy rate of fault samples, and decrease the time spent on the classification.