Expert Systems最新文献

Machine learning and deep learning algorithms have recently been progressively integrated into business intelligence for smart city management. Predicting smart city energy utilisation is crucial for sustainable growth and resource efficiency and become essential due to the growing challenges and demands on raw energy resources. Traditional methods for forecasting energy demand were manual and often yielded poor results. However, advancements in machine learning and deep learning have introduced new approaches to energy management within the context of business intelligence. This research presents an advanced prolonged deep reinforcement learning model incorporating Monte Carlo learning, which is compared against the convolutional neural network approach using a public sector energy dataset, including steel industry energy utilisation. The study evaluates the performance of the proposed prolonged deep reinforcement learning and convolutional neural network models based on key metrics such as accuracy and mean absolute percentage error. Data analysis was conducted using the SPSS tool, encompassing graphical illustrations, group statistics, and independent tabulations. Results indicate that the prolonged deep reinforcement learning model achieved a prediction accuracy of 96.613% and a mean absolute percentage error of 3.387%. Consequently, this efficient prediction model is well-suited for addressing the future energy demands of smart cities through business intelligence.

With increasing focus on sustainable agriculture and AI-enabled solutions, this work proposes AloeVeraNet, a compact deep learning model designed for the efficient and real-time detection of aloe vera leaf diseases on edge devices. The model employs depthwise and pointwise convolutions to achieve a significantly reduced parameter count (289 K) and model size (1.10 MB), enabling deployment in low-resource environments. With 96.09% accuracy, AloeVeraNet sets a new benchmark in classifying aloe vera leaf conditions: healthy, rust-infected and spot-affected, outperforming MobileNetV2, EfficientNetV2-S and VGG16. This sustainable, artificial intelligence (AI)-based solution supports precision agriculture through optimised computation, energy efficiency and local disease monitoring, all without relying on cloud infrastructure, thereby contributing to environmentally responsible farming practices. This study demonstrates the value of integrating AI with sustainable edge computing in creating resilient and inclusive solutions for the agricultural sector.

Industrial robotic manipulators are prone to surface wear and damage, leading to unexpected failures and costly downtimes. Early detection of such defects is crucial for enabling predictive maintenance. This study proposes a vision-based pattern recognition framework that combines convolutional neural networks (CNN) and generative adversarial networks (GANs) to enhance defect detection and tool failure prediction in IoRT-connected environments. The proposed scheme leverages CNN to extract multi-scale visual features from raw images of industrial machine components. Convolutional layers are stacked with varying filter sizes to capture fine-grained surface defects and broader contextual patterns. The pooling layers selectively retain discriminative activations, producing feature embeddings that highlight characteristics such as pitting, scratches, and cracks. This structure allows the network to transform raw pixels into meaningful patterns for reliable classification. To address data scarcity and improve generalisation, the GAN component generates synthetic defect images by simulating real-world variability, including defect shape, background textures and orientation. The adversarial training between the generator and discriminator enhances the realism and diversity of augmented data, which in turn improves the CNN's robustness. Applied to ball screw drive spindle images, the integrated CNN-GAN model achieves 96.7% classification accuracy, with 94% precision, 92% recall and 93% AUC. These results support the system's suitability for predictive maintenance and real-time deployment in smart industrial settings.

News text classification is a technique of classifying news articles into some predefined classes. It helps consumers find news that piques their interest. Due to the growth of internet news content, effective automatic classification systems are required to handle and arrange massive volumes of news articles. Here, the AG's News Corpus (AG News) articles are classified through the hypergraph neural network along with the attention layer and quadratic support vector machine (AGNews_HAL_QSVM). This benchmark dataset was named after the ‘ComeToMyHead’ project by Alberto G. (AG) Leonardo. The dataset was gathered from Kaggle, and the LDA (Latent Dirichlet Allocation) was used to generate the topic-specific data. Every topic will be regarded as a hyperedge in the hypergraph, and each topic's words will be regarded as a hypervertex. A hypergraph convolution neural network with an attention layer is used to extract the corpus' key features. For classification, the collected features are sent into a quadratic support vector machine. A complex deep-learning model has been used to test the proposed model. At an accuracy of 91.2%, the suggested model performs better than the other state-of-the-art algorithms. In order to improve automatic AGNews classification systems, this study presents a practical implementation of the proposed model for organising news content. It propels developments in public discourse, media, personalisation and policy.

Few-shot relation extraction has attracted significant attention due to its potential to identify new relations when training samples are scarce. Previous studies demonstrate that meta-learning techniques can significantly enhance models' adaptability in few-shot scenarios. Most existing meta-learning methods generalize to unseen relations by constructing effective prototype representations for each relation. However, meta-learning techniques often require a large amount of labeled data during the meta-training stage, which contradicts the initial purpose of addressing the issue of insufficient labeled data. Current methods usually fail to model both intra-class similarity and inter-class difference sufficiently, resulting in fuzzy class prototypes when distinguishing between similar but slightly different relations. Therefore, we propose a novel task, low-resource few-shot relation extraction (LR-FSRE), that explores minimizing the usage of labeled data while maximizing the acquisition of task meta-knowledge, and we design a novel graph-based adaptive discrimination network that optimizes relation prototypes in both tasks. Specifically, we effectively represent class prototypes using the topological information between instance and relation-level representation. Then, the relation discrimination network considers the difference between similar classes to further improve the recognition ability of relation classes. Finally, a debiased optimization strategy is employed to optimize the learning processes of task-general and task-specific knowledge. Experimental results show that our proposed framework outperforms the state-of-the-art methods in FSRE and LR-FSRE tasks, and achieves significant improvement in accuracy across challenging cross-domain and similarity relation discrimination scenarios.

Electronic health records have become more prevalent worldwide, and with this, the opportunity for more accurate and automated prediction of health events has grown. Such predictions are crucial for providing preventive and proactive healthcare to patients. Although various advanced methods have been explored, they often fail to fully leverage medical domain knowledge, understand interrelations between diseases and patients comprehensively, and efficiently integrate unstructured clinical notes into predictive models. To address these challenges, we propose the Medical Domain Knowledge Collaborative Graph Learning (MED-CGL) model. MED-CGL incorporates external medical knowledge bases to enhance the predictive power of unstructured clinical notes and extracts learnable features from the MIMIC-III health record dataset using medical domain knowledge and collaborative graph learning. We introduce the Enhanced Medical Knowledge Integration (EMKI) module, which employs a novel attention mechanism to connect clinical notes with disease descriptions precisely. It also enhances the system's performance by integrating medical knowledge from the semantically labelled knowledge-enhanced (SLAKE) dataset during the training phase. Furthermore, our model considers the complexities of unstructured clinical notes, providing a nuanced perspective on the interplay between diseases and patient profiles. Our experiments show that the MED-CGL model exhibited outstanding performance in diagnosis prediction, achieving an F1 score of 27.32%, and in heart failure prediction, where it attained an accuracy of 91.39%. This significant improvement demonstrates the robustness and effectiveness of our model, which is further supported by our in-depth ablation study.

This paper addresses the novel and challenging domain of graph entity dependencies (GEDs) discovery, which aims to identify dependencies in large graphs that are nearly satisfied despite the presence of errors, exceptions and ambiguities in real-world data. We propose a unique error measure specifically designed for GED semantics and innovatively adapts concepts of disagreement and necessary sets to the realm of graph dependencies. Furthermore, we introduce the FastAGEDs+ algorithm, which significantly enhances efficiency in discovering approximate GEDs, employing a depth-first search strategy for optimal candidate space traversal. Incorporating an innovative pruning strategy, FastAGEDs+ efficiently narrows down the search space, significantly reducing computational overhead while maintaining accuracy. Through extensive experimentation on real-world graphs, we demonstrate the feasibility and scalability of our approach, offering substantial improvements in data quality and management practices.

Federated Learning (FL) represents a significant advancement in decentralised machine learning, offering a solution to the privacy concerns associated with traditional centralised approaches. However, a critical limitation of FL arises in the presence of Non-Independent and Identically Distributed (non-IID) data, which is common in real-world scenarios. Traditional FL algorithms, such as Federated Averaging (FedAvg), tend to underperform when faced with data heterogeneity across participating clients. To address this challenge, we propose CalibSNN, a method that combines calibration re-sampling with Soft Nearest Neighbour (SNN) loss to mitigate the bias and variance introduced by uneven data distributions. Calibration aligns local data distributions with global statistics, while SNN loss improves feature representations across heterogeneous clients. Through extensive experiments on diverse datasets and non-IID conditions, we demonstrate that CalibSNN significantly outperforms state-of-the-art baselines, offering a robust solution to the challenges of non-IID data in FL.