This study introduces a data augmentation technique based on generative adversarial networks (GANs) to improve the accuracy of day-ahead wind power predictions. To address the peculiarities of abrupt weather data, we propose a novel method for detecting mutation rates (MR) and local mutation rates (LMR). By analyzing historical data, we curated datasets that met specific mutation rate criteria. These transformed wind speed datasets were used as training instances, and using GAN-based methodologies, we generated a series of augmented training sets. The enriched dataset was then used to train the wind power prediction model, and the resulting prediction results were meticulously evaluated. Our empirical findings clearly demonstrate a significant improvement in the accuracy of day-ahead wind power prediction due to the proposed data augmentation approach. A comparative analysis with traditional methods showed an approximate 5% increase in monthly average prediction accuracy. This highlights the potential of leveraging mutated wind speed data and GAN-based techniques for data augmentation, leading to improved accuracy and reliability in wind power predictions. In conclusion, this paper presents a robust data augmentation method for wind power prediction, contributing to the potential enhancement of day-ahead prediction accuracy. Future research could explore additional mutation rate detection methods and strategies to further enhance GAN models, thereby amplifying the effectiveness of wind power prediction.

Due to the increasing proportion of wind power connected to the grid, day-ahead wind power prediction plays a more and more important role in the operation of the power system. This paper proposes a day-ahead wind power short-term prediction model based on deep learning (DWT_AE_BiLSTM). Firstly, discrete wavelet transform (DWT) is used to denoise the data, then an autoencoder (AE) technology is used to extract the data features, and finally, bidirectional long short-term memory (BiLSTM) is used for prediction. To verify the effectiveness of the proposed DWT_AE_BiLSTM model, we studied three different power stations and compared their performance with the shallow neural network model. Experimental analysis shows that this model is more competitive in forecasting accuracy and stability. Compared with the BP model, the proposed model has increased by 3.86%, 3.22% and 3.42% in three wind farms, respectively.

The evaluation of the effects of artificial precipitation enhancement remains one of the most important and challenging issues in the fields of meteorology. Rainfall is the most important evaluation metric for artificial precipitation enhancement, which is mainly achieved through physics-based models that simulate physical phenomena and data-driven statistical models. The series of effect evaluation methods requires the selection of a comparison area for effect comparison, and idealized assumptions and simplifications have been made for the actual cloud precipitation process, leading to unreliable quantitative evaluation results of artificial precipitation effects. This paper proposes a deep learning-based method (UNET-GRU) to quantitatively evaluate the effect of artificial rainfall. By comparing the residual values obtained from inverting the natural evolution grid rainfall of the same area under the same artificial rainfall conditions with the actual rainfall amount after artificial rainfall operations, the effect of artificial rainfall can be quantitatively evaluated, effectively solving the problem of quantitative evaluation of artificial precipitation effects. Wuhan and Shiyan in China are selected to represent typical plains and mountainous areas, respectively, and the method is evaluated using 6-min resolution radar weather data from 2017 to 2020. During the experiment, we utilized the UNET-GRU algorithm and developed separate algorithms for comparison against common persistent baselines (i.e., the next-time data of the training data). The prediction of mean squared error (MSE) for these three algorithms was significantly lower than that of the baseline data. Moreover, the indicators for these algorithms were excellent, further demonstrating their efficacy. In addition, the residual results of the estimated 7-h grid rainfall were compared with the actual recorded rainfall to evaluate the effectiveness of artificial precipitation. The results showed that the estimated rainfall was consistent with the recorded precipitation for that year, indicating that deep learning methods can be successfully used to evaluate the impact of artificial precipitation. The results demonstrate that this method improves the accuracy of effect evaluation and enhances the generalization ability of the evaluation scheme.

Hierarchical Federated Learning (HFL) significantly enhances communication efficiency and device participation, while improving personalized learning outcomes. In this framework, incentive mechanisms are crucial as they ensure that devices actively participate and make genuine contributions. However, existing incentive mechanisms struggle to effectively address the issue of unreliable devices, which may negatively impact model training due to malicious behavior or faults, leading to low-quality updates or even failure of the global model. Additionally, participants’ strategic behaviors and device heterogeneity can further diminish the effectiveness of these mechanisms. To tackle these challenges, this paper proposes a Reputation-Aware Incentive Mechanism (RAIM) aimed at optimizing node cooperation within HFL and enhancing overall system performance. Specifically, we first evaluate the reputation value of end devices based on their training quality and historical records, which can identify and defend against malicious data attacks. Participants’ reputations are maintained through a consortium blockchain, thereby ensuring transparency and fairness. Next, we model the interaction of HFL as a three-stage Stackelberg game to address hierarchical decision-making processes, and also prove that there is a unique Stackelberg equilibrium, derived through cautiously proposed algorithms. Since the existing equilibrium may not be optimal, we further design optimal server selection algorithm to motivate high-reputation and low-cost devices to participate in training, while maximizing both system performance and social utility. Finally, extensive experiments using both synthetic and real datasets show that our RAIM outperforms state-of-the-art baseline methods.

Personalized recommendation for online learning courses stands as a critical research topic in educational technology, where algorithmic performance directly impacts learning efficiency and user experience. To address the limitations of existing studies in multimodal heterogeneous data fusion and high-order relationship modeling, this research proposes a Heterogeneous Hypergraph and Attention-based Online Course Recommendation (HHAOCR) algorithm. By constructing a heterogeneous hypergraph structure encompassing three entity types (students, instructors, and courses), we innovatively designed hypergraph convolution operators to achieve bidirectional vertex-hyperedge information aggregation, integrated with a dynamic attention mechanism to quantify important differences among entities. The method establishes computational frameworks for hyperedge-vertex coefficient matrices and inter-hyperedge attention scores, effectively capturing high-order nonlinear correlations within multimodal heterogeneous data, while employing temporal attention units to track the evolution of user preferences. Experimental results on the MOOCCube dataset demonstrate that the proposed algorithm achieves significant improvements in NDCG@15 and F1-Score@15 metrics compared to TP-GNN (enhanced by 0.0699 and 0.0907) and IRS-GCNet (enhanced by 0.0808 and 0.0999). This work provides a scalable solution for multisource heterogeneous data fusion and precise recommendation for online education platforms.

To address the problem of inaccurate matching between personalized exercise recommendations and learners’ mastery of knowledge concepts/learning abilities, we propose the Dynamic Multidimensional Memory Augmented knowledge tracing model (DMMA). This model integrates a dynamic key-value memory neural network with the Ebbinghaus Forgetting Curve. By incorporating time decay factors and knowledge concept mastery speed factors, it dynamically adjusts knowledge update intensity, effectively resolving the insufficient personalized recommendation capabilities of traditional models. Experimental validation demonstrates its effectiveness: on Algebra 2006–2007, DMMA achieves 82% accuracy, outperforming CRDP-KT by 6%, while maintaining 53–55% accuracy for cold-start users (0–5 interactions), which is 25% higher than CoKT. The model’s integration of the Ebbinghaus forgetting curve and K-means-based concept classification enhances adaptability. Genetic algorithm optimization yields a diversity score of 0.79, with 18% higher 30-day knowledge retention. The FastDTW–Sigmoid hybrid similarity calculation (weight transition 0.27–0.88) ensures smooth cold-start adaptation, while novelty metrics reach 0.65 via random-forest-driven prediction. Ablation studies confirm component necessity: removing time decay factors reduces accuracy by 2.2%. These results validate DMMA’s superior performance in personalized education.