Traditional Learning Methods Research Articles

Remaining useful life prediction with spatio-temporal graph transform and weakly supervised adversarial network: An application in power components

Remaining useful life (RUL) prediction is crucial in improving the reliability of insulated gate bipolar transistors (IGBTs) and has attracted widespread attention all over the world. Owing to the heterogeneity and complex internal connectivity of the IGBTs, accurate RUL prediction by traditional deep learning (DL) methods remains challenging. Herein, a novel prognostic framework is proposed for their RUL prediction. In brief, the non-smooth spike voltage of IGBTs is transformed into a graph matrix that contains rich correlation information. Then, spatio-temporal fusion graph network (STFGN) is proposed to realize RUL prediction. It contains spatial-scale and temporal-scale extractors, where the former one based on GraphSage can obtain unique IGBTs data structure information, while the latter based on bidirectional long short-term memory (BiLSTM) captures the time-series degradation features of IGBTs. Moreover, the weakly supervised adversarial learning strategy is considered to realize cross-domain prediction using scarce labeled data, which enhance the positive transfer. Four transfer tasks under variable working conditions are considered to prove the proposed method. The experimental results demonstrate that the average SF and RMSE metrics of the proposed method are 0.429 and 0.108, which is superior to other related methods.

Design Thinking Engineering Learning on Cloud Ecosystem Model to Enhance Digital Intelligence for Undergraduate Student

This article aims to present a new educational model that integrates engineering learning with design thinking via a cloud ecosystem to enhance digital intelligence for undergraduate students. This approach utilizes user-centered problem-solving to foster innovation and collaborative learning. The educational environment can adapt to 21st-century challenges by leveraging technology and empowering students to engage confidently with various digital tools and platforms. This model promotes digital and emotional intelligence in digital interactions, essential for modern workplace success. The effectiveness of this learning model has been assessed through structured evaluations that highlight its potential to transform traditional learning methods and prepare students for the technological landscape of the future.

Explainable ensemble deep learning-based model for brain tumor detection and classification

AbstractBrain tumors are very dangerous as they cause death. A lot of people die every year because of brain tumors. Therefore, accurate classification and detection in the early stages can help in recovery. Various deep learning techniques have achieved good results in brain tumor classification. The traditional deep learning methods and training the neural network from scratch are time-consuming and can last for weeks of training. Therefore, in this work, we proposed an ensemble approach depending on transfer learning that utilizes pre-trained models of DenseNet121 and InceptionV3 to detect three forms of brain tumors: meningioma, glioma, and pituitary. While developing the ensemble model, some changes were made to the architecture of pre-trained models by replacing their classifiers (fully connected and SoftMax layers) with a new classifier to adopt the recent task. In addition, gradient-weighted class activation maps (Grad-CAM) are an explainable model to verify results and achieve high confidence. The suggested model was validated using a publicly available dataset and achieved 99.02% accuracy, 98.75% precision, 98.98% recall, and a 98.86% F1 score. The suggested approach outperformed others in detecting and classifying brain tumor MRI data, and verifying results using the explainable model achieved a high degree of trust.

Hypergraph Neural Network for Multimodal Depression Recognition

Deep learning-based approaches for automatic depression recognition offer advantages of low cost and high efficiency. However, depression symptoms are challenging to detect and vary significantly between individuals. Traditional deep learning methods often struggle to capture and model these nuanced features effectively, leading to lower recognition accuracy. This paper introduces a novel multimodal depression recognition method, HYNMDR, which utilizes hypergraphs to represent the complex, high-order relationships among patients with depression. HYNMDR comprises two primary components: a temporal embedding module and a hypergraph classification module. The temporal embedding module employs a temporal convolutional network and a negative sampling loss function based on Euclidean distance to extract feature embeddings from unimodal and cross-modal long-time series data. To capture the unique ways in which depression may manifest in certain feature elements, the hypergraph classification module introduces a threshold segmentation-based hyperedge construction method. This method is the first attempt to apply hypergraph neural networks to multimodal depression recognition. Experimental evaluations on the DAIC-WOZ and E-DAIC datasets demonstrate that HYNMDR outperforms existing methods in automatic depression monitoring, achieving an F1 score of 91.1% and an accuracy of 94.0%.

Intelligent Prestack Multi-trace Seismic Inversion Constrained by Probabilistic Geological Information

Seismic inversion uses observed data including well-logging and seismic data to infer rock parameters. However, it still suffers from non-unique solutions. The non-uniqueness increase when extrapolating away from well location. To mitigate this effect, we incorporate geological information into a deep learning (DL) framework in the form of generated probabilistic labels. We transform geological information into reservoir heterogeneity (RH) weights and geological pattern constraint (GPC) weights to generate probabilistic labels. The RH weights are used to capture the variability in accuracy during the spatial extrapolation of well data. The RH weights are used to balance the constraint weights between well and seismic data in the DL model. The GPC weights are used to characterize the changes of geological patterns from well locations to non-well locations. Based on variations in local prestack seismic waveforms, the DL model learns extrapolation patterns of well data with similar geological patterns. By combining the rock parameters derived from well logs with these two types of weights, we can generate a series of probability labels. Two types of geological information and well-logging data are deeply integrated into a supervised 2D-ResUNet framework. In this way, we develop a novel DL-based prestack multi-trace seismic inversion method. We validate our method on a 3D synthetic model and a 3D field dataset. The results show that our method exhibits great potential for characterizing the spatial variation of subsurface rocks and outperforms traditional deep learning methods.

Path Planning for Robots Combined with Zero-Shot and Hierarchical Reinforcement Learning in Novel Environments

Path planning for robots based on reinforcement learning encounters challenges in integrating semantic information about environments into the training process. In unseen or complex environmental information, agents often perform sub-optimally and require more training time. In response to these challenges, this manuscript pioneers a framework integrating zero-shot learning combined with hierarchical reinforcement learning to enhance agent decision-making in complex environments. Zero-shot learning enables agents to infer correct actions for previously unseen objects or situations based on learned semantic associations. Subsequently, the path planning component utilizes hierarchical reinforcement learning with adaptive replay buffer, directed by the insights gained from zero-shot learning, to make decisions effectively. Two parts are trained separately, so zero-shot learning is available in different and unseen environments. Through simulation experiments, we compare the traditional hierarchical reinforcement learning method with the proposed method. The results prove that this structure can make full use of environmental information to generalize across unseen environments and plan collision-free paths.

Methodology for Object-Level Change Detection in Post-Earthquake Building Damage Assessment Based on Remote Sensing Images: OCD-BDA

Remote sensing and computer vision technologies are increasingly leveraged for rapid post-disaster building damage assessment, becoming a crucial and practical approach. In this context, the accuracy of employing various AI models in pixel-level change detection methods is significantly dependent on the consistency between pre- and post-disaster building images, particularly regarding variations in resolution, viewing angle, and lighting conditions; in object-level feature recognition methods, the low richness of semantic details of damaged buildings in images leads to a poor detection accuracy. This paper proposes a novel method, OCD-BDA (Object-Level Change Detection for Post-Disaster Building Damage Assessment), as an alternative to pixel-level change detection and object-level feature recognition methods. Inspired by human cognitive processes, this method incorporates three key steps: an efficient sample acquisition for object localization, labeling via HGC (Hierarchical and Gaussian Clustering), and model training and prediction for classification. Furthermore, this study establishes a change detection dataset based on Google Earth imagery of regions in Hatay Province before and after the Turkish earthquake. This dataset is characterized by pixel inconsistency and significant differences in photographic angles and lighting conditions between pre- and post-disaster images, making it a valuable test dataset for other studies. As a result, in the experiments of comparative generalization capabilities, OCD-BDA demonstrated a significant improvement, achieving an accuracy of 71%, which is twice that of the second-ranking method. Moreover, OCD-BDA exhibits superior performance in tasks with small sample amounts and rapid training time. With only 1% of the training samples, it achieves a prediction accuracy exceeding that of traditional transfer learning methods with 60% of samples. Additionally, it completes assessments across a large disaster area (450 km²) with 93% accuracy in under 23 min.

Implementation of the 2013 Curriculum in Integrated Thematic Learning in Elementary Schools: Strategies, Challenges, and its Impact on Learning

Education is the process of imparting knowledge, character development, and producing a more harmonious social order and way of life, including through the development of students' intellectual capital, skills, and intellectual development. One of the innovations brought by the introduction of the 2013 curriculum is the integrative theme learning model. To provide meaningful educational experiences to students, integrative thematic learning connects various disciplines and materials through themes. The 2013 curriculum states that in order to place students as the center of the learning process, teachers must play an active role as facilitators and motivators of learning. This article aims to explain how the 2013 curriculum is applied to integrative thematic learning, and its impact on the teaching and learning process in elementary schools. The implementation of the curriculum in integrative thematic learning shows several challenges, such as difficulties in determining assessment or evaluation criteria for students, teachers' lack of understanding in compiling thematic integrative lesson plans, limited teaching aids, and the use of traditional learning methods that are less varied. The proposed solution is training and education for teachers, principals, and supervisors to acquire skills and knowledge in implementing the curriculum according to students' needs. By implementing a variety of learning methods, it is hoped that it can increase student motivation and independence in the learning process.

Improving ranking-based question answering with weak supervision for low-resource Qur’anic texts

This work tackles the challenge of ranking-based machine reading comprehension (MRC), where a question answering (QA) system generates a ranked list of relevant answers for each question instead of simply extracting a single answer. We highlight the limitations of traditional learning methods in this setting, particularly under limited training data. To address these issues, we propose a novel ranking-inspired learning method that focuses on ranking multiple answer spans instead of single answer extraction. This method leverages lexical overlap as weak supervision to guide the ranking process. We evaluate our approach on the Qur’an Reading Comprehension Dataset (QRCD), a low-resource Arabic dataset over the Holy Qur’an. We employ transfer learning with external resources to fine-tune various transformer-based models, mitigating the low-resource challenge. Experimental results demonstrate that our proposed method significantly outperforms standard mechanisms across different models. Furthermore, we show its better alignment with the ranking-based MRC task and the effectiveness of external resources for this low-resource dataset. Our best performing model achieves a state-of-the-art partial Reciprocal Rank (pRR) score of 63.82%, surpassing the previous best-known score of 58.60%. To foster further research, we release code [GitHub repository:github.com/mohammed-elkomy/weakly-supervised-mrc], trained models, and detailed experiments to the community.

Offline Reinforcement Learning for Adaptive Control in Manufacturing Processes: A Press Hardening Case Study

Abstract This paper explores the application of offline reinforcement learning in batch manufacturing, with a specific focus on press hardening processes. Offline reinforcement learning presents a viable alternative to traditional control and reinforcement learning methods, which often rely on impractical real-world interactions or complex simulations and iterative adjustments to bridge the gap between simulated and real-world environments. We demonstrate how offline reinforcement learning can improve control policies by leveraging existing data, thereby streamlining the training pipeline and reducing reliance on high-fidelity simulators. Our study evaluates the impact of varying data exploration rates by creating five datasets with exploration rates ranging from ε=0 to ε=0.8. Using the conservative Q-learning algorithm, we train and assess policies against both a dynamic baseline and a static industry-standard policy. The results indicate that while offline reinforcement learning effectively refines behavior policies and enhances supervised learning methods, its effectiveness is heavily dependent on the quality and exploratory nature of the initial behavior policy.

Action research plan: a methodology to examine the impact of artificial intelligence (AI) on the cognitive abilities of university students

This research suggests a methodology to examine the effectiveness Artificial Intelligence (AI) on the cognitive abilities of college students so that future researchers can utilize this experimental project to focus on how AI-powered Intelligent Tutoring Systems (ITSs) affect learning outcomes. As AI continues to revolutionize all walks of life, so has its integration with education. ITS is a key application of AI in education, providing a personalized learning experience by analyzing student data and tailoring instructional materials accordingly. This provides a research program to study the effectiveness of ITS in improving cognitive skills such as memory, critical thinking, and problem solving abilities in college students. The research project combined quantitative and qualitative research methods including surveys, pre-tests, post-tests, and in-depth interviews to assess the cognitive differences between students using ITS and those using traditional learning methods. The research project also recognized potential challenges, such as dependence on technology and the risk of increased educational inequality. The action research program concluded by advocating for a balanced integration of AI into education, highlighting the need for ongoing research to optimize its use across different stages of education and to ensure equitable access to AI-powered tools.

Research on positive and negative sample sampling strategies in contrastive learning

Abstract. Contrastive learning is an important technique in the field of machine learning for learning data representations. In the field of self-supervised visual representation learning, the strategy of positive and negative sample selection is key to improving the efficiency and effectiveness of model learning. Traditional self-supervised learning methods often employ random sampling strategies to select positive and negative samples, but this approach may result in uneven sample quality when dealing with complex datasets, thereby affecting the learning outcomes. To alleviate this problem, this study is dedicated to exploring more effective strategies for positive and negative sample selection and processing to optimize the self-supervised learning process. To this end, we propose an improved method of self-supervised learning called contrastive learning with enhanced diversity. On the one hand, this method utilizes the weight parameters of the DINO pre-trained model to initialize the feature extraction network of SimCLR, providing more accurate calculations of feature similarity. On the other hand, by setting a threshold on the feature similarity matrix and penalizing (subtracting 0.5 from) similarity scores that do not exceed this threshold, we reduce the excessive impact of high similarity scores on model training, thereby helping the model better distinguish between positive and negative samples. In downstream image classification tasks, we conducted detailed evaluations of the improved model, specifically including fine-tuning and linear evaluation aspects. Experiments show that the proposed approach improves the performance of the loss functions and improves the accuracy of the proposed SimCLR model.

Comparison Distributed and Parallel Machine Learning: Evidence from Models, Principles and Application Scenarios

Abstract. Contemporarily, the demand for processing large-scale data has been rapidly increasing, prompting continuous advancements in the field of machine learning. This study examines the state of parallel and distributed machine learning at present, both of which aim to enhance computational efficiency when dealing with large datasets and demonstrate promising applications across various domains. As the scale of data escalates, traditional single-machine learning methods are becoming increasingly inadequate, which lead to the emergence of parallel and distributed machine learning. These approaches enable substantial computations to be performed more efficiently through the collaboration of multi-core CPUs or multiple computing nodes. This research conducts an in-depth analysis of the inherent challenges associated with these methods, including data transmission latency, synchronization requirements, and user privacy concerns. Ultimately, this research emphasizes the tremendous potential of both methods for future applications, a potential that is bolstered by ongoing advancements in hardware and algorithm optimization. These results provide valuable insights for practitioners in the field and offers guidance for future research directions.

Pathways to motivation and learning in English language education: a case of two youths engaging in multiliteracies

Purpose This study investigates the effect of digital multimodal composing (DMC) on student motivation and literacy in English as a Foreign Language (EFL) contexts. It explores how students engaged in creating digital animations to respond to literary texts and the implications for integrating DMC into traditional literacy practices. Grounded in social semiotic theory and Self-Determination Theory, this study aims to understand how digital composing influences student engagement, motivation and literacy development. Design/methodology/approach Qualitative data were gathered over a six-month period using a combination of student surveys, semi-structured interviews, reflection logs and student-produced scripts/animations. This study focused on two university students who engaged in multimodal projects as part of their extracurricular EFL activities, examining their processes and reflections. Findings Students demonstrated strong intrinsic motivation during DMC processes, largely due to the social affordances provided by digital multimodal composing. By combining traditional literacy skills with digital practices, students experienced enhanced creative expression, enhanced meaning-making and a deeper understanding of literary texts through the production and sharing of animated videos. This process promoted both emotional and cognitive engagement, emphasizing the importance of teacher guidance and digital tools in creating a more dynamic and supportive learning environment. Practical implications This study outlines a practical approach to DMC through a structured process: selecting and reading a book, proposing a change as if requesting it from the author, creating and sharing a video on YouTube to explain the change and reflecting on the experience. This approach can guide educators in designing activities that deepen student engagement, enhance critical thinking and integrate digital tools with traditional learning methods. Originality/value This study contributes to the understanding of how digital technologies can transform traditional literacy practices, offering insights into the benefits of multimodal composition for language learning and motivation in EFL contexts. It provides practical examples of how students and teachers can leverage digital tools to enhance literacy education and prepare students for 21st-century challenges.

Virtual reality in medical education: Effectiveness of Immersive Virtual Anatomy Laboratory (IVAL) compared to traditional learning approaches

Immersive Virtual Anatomy Laboratory (IVAL) is an innovative learning tool that combines virtual reality and serious games elements to enhance anatomy education. This experimental study compares IVAL with traditional learning methods in terms of educational effectiveness and user acceptance. An experimental design was implemented with 120 undergraduate health-science students, randomly assigned to two groups: an experimental group using IVAL, and a control group following traditional learning methods. Data collection focused on quantitative measures such as pretest and posttest vocabulary assessment scores and task completion times, alongside qualitative measures obtained through a user experience questionnaire. This study utilizes the Technology Acceptance Model (TAM), incorporating variables such as Perceived Usefulness and Perceived Ease of Use. Results revealed significant improvements in the experimental group, with a 55.95% increase in vocabulary scores and an 18.75% reduction in task completion times compared to the control group. Qualitative data indicated that IVAL users reported greater Perceived Usefulness of the technology, improved Perceived Ease of Use, a more positive Attitude Towards Using IVAL, and stronger Behavioral Intention to continue using IVAL for anatomy learning. This study demonstrates that the integration of immersive virtual reality in the IVAL approach offers a promising method to enhance anatomy education. The findings provide insights into the effectiveness of immersive learning environments in improving learning outcomes and user acceptance. While further research is needed to explore long-term effects, this innovative approach not only enhances the effectiveness and enjoyment of anatomy learning but also provides valuable data on optimizing educational technology for improved learning outcomes.

The Difficulties That Faced Libyan Students in Learning English Vocabulary and Understanding Academic Texts

This study examines Libyan students' difficulties in acquiring the English language and comprehending academic writing. The research delineates particular vocabulary-related challenges and their effects on reading comprehension and academic achievement. A quantitative methodology was employed, utilizing a questionnaire administered to 32 students from the English department at Aljufra University. The results demonstrate that students face considerable difficulties in enhancing their vocabulary, employing words correctly, and comprehending phrasal verbs. The challenges are frequently intensified by cultural and linguistic disparities between Arabic and English, impeding students' ability to identify equivalent concepts in both languages. It also reveals that many students rely on traditional learning methods, such as dictionaries, flashcards, and repetition. While these strategies are somewhat effective, they are insufficient for mastering the complexities of academic English. The research highlights the necessity for educational institutions to provide additional support, including targeted vocabulary instruction, workshops, and one-on-one tutoring, to enhance students' academic writing and comprehension skills. Over 60% of the participants expressed the need for more excellent institutional support in academic writing, emphasizing the importance of structured resources to improve their learning outcomes. The results further emphasize the challenges associated with phrasal verbs and complex vocabulary, which remain common obstacles for non-native speakers. Addressing these areas through culturally responsive teaching techniques and providing real-world examples could make learning more relatable for students. The research underlines the importance of a comprehensive approach to vocabulary learning that incorporates both traditional and modern methods. Educators' focused effort to provide vocabulary support, academic assistance, and culturally relevant instruction is essential for improving Libyan students' vocabulary, enhancing their comprehension of scholarly texts, and ultimately enabling their educational success.

Assessing traditional and machine learning methods to smooth and impute device-based body condition score throughout the lactation in dairy cows

Regular monitoring of body condition score (BCS) changes during lactation is an essential management tool in dairy cattle; however, the current BCS measurements are often discontinuous and unevenly spaced in time. The imputation of BCS values is useful for two main reasons: i) achieving completeness of data is necessary to be able to relate BCS to other traits (e.g. milk yield and milk composition) that have been routinely recorded at different times and with a different frequency, and ii) having expected BCS values provides the possibility to trigger early warnings for animals with certain unexpected conditions. The contribution of this study was to propose and evaluate potential methods useful to smooth and impute device-based BCS values recorded during lactation in dairy cattle. In total, 26,207 BCS records were collected from 3,038 cows (9,199 and 14,462 BCS records on 1,546 Holstein and 1,211 Montbéliarde cows respectively, and the rest corresponded to other minority cattle breeds). Six methods were evaluated to predict BCS values: the traditional methods of test interval method (TIM), and multiple-trait procedure (MTP), and the machine learning (ML) methods of multi-layer perceptron (MLP), Elman network (Elman), long-short term memories (LSTM) and bi-directional LSTM (BiLSTM). The performance of each method was evaluated by a hold-out validation approach using statistics of the root mean squared error (RMSE) and Pearson correlation (r). TIM, MTP, MLP, and BiLSTM were assessed for the imputation of intermediate missing values, while MTP, Elman, and LSTM were evaluated for the forecasting of future BCS values. Regarding the machine learning methods, BiLSTM demonstrated the best performance for the intermediate value imputation task (RMSE = 0.295, r = 0.845), while LSTM demonstrated the best performance for the future value forecasting task (RMSE = 0.356, r = 0.751). Among the methods evaluated, MTP showed the best performance for imputation of intermediate missing values in terms of RMSE (0.288) and r (0.856). MTP also achieved the best performance for forecasting of future BCS values in terms of RMSE (0.348) and r (0.760). This study demonstrates the ability of MTP and machine learning methods to impute missing BCS data and provides a cost-effective solution for the application area.

The Padlet Media Design Based on Bima Local Literature to Improve Student Writing Skills

This research aims to improve students' writing skills through the use of Padlet media which is integrated with local wisdom of Bima regional literature. The background of this research is based on the low writing ability of students caused by monotonous traditional learning methods and the lack of use of interactive media. This research uses the Research and Development (R&D) method with the ADDIE (Analysis, Design, Development, Implementation, and Evaluation) development model. Local literature, such as folklore and poetry, is integrated into Padlet's platform as teaching materials to provide relevant cultural context for students. The participants in this study are 30 students of the Indonesian Language and Literature Education Study Program at STKIP Harapan Bima. The results of the study showed a significant improvement in students' writing skills after the use of Padlet based on local wisdom. The average writing score of students increased by 25%, especially in the aspects of cohesion, coherence, and idea development. In addition, students also reported increased motivation and active involvement in writing learning through Padlet's collaborative features, such as writing sharing and feedback. However, there are several technical obstacles faced, such as limited internet access and lack of digital literacy among students. The results of this study have implications that the integration of digital technology and local wisdom can be an innovative strategy in improving students' writing skills, as well as fostering their cultural awareness.

Enhanced landslide susceptibility mapping in data-scarce regions via unsupervised few-shot learning

Given the critical need to assess landslide hazards, producing landslide susceptibility map (LSM) in regions with scarce historical landslide inventories poses significant challenges. This study introduces a novel landslide susceptibility assessment framework that combines unsupervised learning strategies with few-shot learning methods to increase the accuracy of LSM in these areas. The framework has been practically validated in a representative geological disaster-prone area along the West-East Gas Pipeline in Shaanxi Province, China. We employed three advanced few-shot learning models: a support vector machine, meta-learning, and transfer learning. These models implement feature representation learning for weakly correlated influencing factors through an unsupervised approach, thereby constructing an effective landslide susceptibility assessment model. We compared traditional learning methods and used the receiver operating characteristic (ROC) curve and SHAP values to quantify the effectiveness of the models. The results indicate that the meta-learning algorithm outperforms both the SVM and transfer learning in areas with limited landslide data. The integration of unsupervised strategies significantly improves performance, achieving area under the curve (AUC) values of 0.9385 and 0.9861, respectively. Compared with using meta-learning alone, incorporating unsupervised learning strategies increased the AUC by 4.76%, enhancing both the predictive power of the model and the interpretability of the features. Meta-learning under unsupervised conditions effectively mitigates the evaluation difficulties caused by insufficient landslide records, providing a viable path and empirical evidence for performance improvement in similar data- scarce regions worldwide.

A rolling bearing fault diagnosis method for imbalanced data based on multi-scale self-attention mechanism and novel loss function

Deep learning methods are widely used in the field of rolling bearing fault diagnosis and produce good results when faced with datasets with roughly equal numbers of normal and faulty samples. However, real-world data often has a serious imbalance, with the number of fault samples being significantly less than the number of normal samples. This dataset imbalance challenges the performance of traditional deep learning methods. To address this problem, this paper proposes an efficient imbalanced data rolling bearing fault diagnosis method. The method consists of two parts: a deep learning network based on a multi-scale self-attention mechanism and a novel loss function. In terms of the deep learning network, firstly, the one-dimensional vibration signal is converted into a two-dimensional image through the Gramian angular field. This conversion maximises the inherent feature extraction capability of the network. Subsequently, the multi-scale learning capability of the network is enhanced by implementing different expansion rates for the head of the multi-scale self-attention mechanism. This nuanced approach allows the network to capture the underlying information more efficiently. Finally, the inclusion of Ghost bottlenecks and feature pyramid networks (FPNs) helps to optimise network efficiency and improve generalisation performance. A novel loss function is also proposed to make the method more suitable for imbalanced data. During the training process, the classification of samples in each class is analysed using the recall metric of imbalanced classification and the real-time recall is used as a weight to weaken the dominance of the majority class. This weighting ensures the adaptability of the method to imbalanced datasets. The proposed method is evaluated using rolling bearing datasets from Case Western Reserve University, USA, and Southeast University, China. Comparison results with other state-of-the-art deep learning methods show that the proposed method has a robust performance when dealing with imbalanced data.