Intelligent Systems with Applications最新文献_第2页

Maximizing the spread of information through content optimization 通过优化内容最大限度地传播信息

Intelligent Systems with Applications

Pub Date : 2024-10-01 DOI: 10.1016/j.iswa.2024.200448

Lei Lin , Yihua Du , Shibo Zhao , Wenkang Jiang , Qirui Tang , Li Xu

As data-driven prediction models advance, an increasing number of people are enjoying news personalized to their interests. The primary problem such recommendation models solve is to precisely match information with users and, in so doing, ensure that news spreads with greater efficiency. However, these techniques only help the media platform; they do not help those who produce the news. Hence, we devised a propagation framework based on a human-in-the-loop simulation that helps content authors maximize the spread of their messages through social networks. The framework works by acting on feedback provided by the simulation model. Additionally, the spread of information is formulated as a multi-objective optimization problem in which propagation is data-driven and simulated with machine learning techniques that leverage data on the historical behaviors of users. We additionally describe an implementation for this framework as an example of how the framework might be used in real life. On the practical side, the implementation uses text data from a blog to simulate the message's propagation, while, from a technical point of view, the multi-objective optimization problem is divided into an information retrieval problem and an integer programming problem, the results of which are fed back into the content editor as content operation strategies. A case study with the Sina Weibo microblog site not only validates the framework but also provides practitioners with insights into how to maximize the spread of information through social networking platforms. The results show that the proposed propagation framework is capable of increasing retweets by 7.9575 %. As an interesting aside, our experiments also show that the Weibo retweet lottery is both popular and a highly effective mechanism for increasing reposts.

随着数据驱动的预测模型的发展，越来越多的人开始享受符合自己兴趣的个性化新闻。此类推荐模型解决的主要问题是将信息与用户进行精确匹配，从而确保新闻传播的效率更高。然而，这些技术只能帮助媒体平台，却无法帮助新闻生产者。因此，我们设计了一个基于 "人在回路中 "模拟的传播框架，帮助内容作者最大限度地通过社交网络传播信息。该框架根据仿真模型提供的反馈采取行动。此外，信息传播被表述为一个多目标优化问题，其中传播是由数据驱动的，并通过机器学习技术利用用户历史行为数据进行模拟。此外，我们还介绍了该框架的实现方法，以此为例说明如何在现实生活中使用该框架。在实际应用方面，该实现使用博客中的文本数据来模拟信息的传播，而从技术角度来看，多目标优化问题分为信息检索问题和整数编程问题，其结果作为内容操作策略反馈给内容编辑器。通过对新浪微博网站的案例研究，不仅验证了该框架，还为实践者提供了如何通过社交网络平台实现信息传播最大化的见解。结果表明，所提出的传播框架能够将转发量提高 7.9575%。有趣的是，我们的实验还表明，微博转发抽奖既受欢迎，又是一种非常有效的增加转发的机制。

{"title":"Maximizing the spread of information through content optimization","authors":"Lei Lin , Yihua Du , Shibo Zhao , Wenkang Jiang , Qirui Tang , Li Xu","doi":"10.1016/j.iswa.2024.200448","DOIUrl":"10.1016/j.iswa.2024.200448","url":null,"abstract":"<div><div>As data-driven prediction models advance, an increasing number of people are enjoying news personalized to their interests. The primary problem such recommendation models solve is to precisely match information with users and, in so doing, ensure that news spreads with greater efficiency. However, these techniques only help the media platform; they do not help those who produce the news. Hence, we devised a propagation framework based on a human-in-the-loop simulation that helps content authors maximize the spread of their messages through social networks. The framework works by acting on feedback provided by the simulation model. Additionally, the spread of information is formulated as a multi-objective optimization problem in which propagation is data-driven and simulated with machine learning techniques that leverage data on the historical behaviors of users. We additionally describe an implementation for this framework as an example of how the framework might be used in real life. On the practical side, the implementation uses text data from a blog to simulate the message's propagation, while, from a technical point of view, the multi-objective optimization problem is divided into an information retrieval problem and an integer programming problem, the results of which are fed back into the content editor as content operation strategies. A case study with the Sina Weibo microblog site not only validates the framework but also provides practitioners with insights into how to maximize the spread of information through social networking platforms. The results show that the proposed propagation framework is capable of increasing retweets by 7.9575 %. As an interesting aside, our experiments also show that the Weibo retweet lottery is both popular and a highly effective mechanism for increasing reposts.</div></div>","PeriodicalId":100684,"journal":{"name":"Intelligent Systems with Applications","volume":"24 ","pages":"Article 200448"},"PeriodicalIF":0.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142526532","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Masked face image segmentation using a multilevel threshold with a hybrid fitness function 使用混合拟合函数的多级阈值进行屏蔽人脸图像分割

Intelligent Systems with Applications

Pub Date : 2024-09-29 DOI: 10.1016/j.iswa.2024.200445

Nada AbdElFattah Ibrahim , Ehab R. Mohamed , Hanaa M. Hamza , Yousef S. Alsahafi , Khalid M. Hosny

Masked face segmentation tasks have become significantly more difficult due to the increasing use of face masks. On the other hand, the forehead, eyebrows, and eye regions are usually visible and reveal vital information. This exposed area of the face has been segmented and trusted to be used in real life for various applications, such as security, healthcare education, and projects in smart cities. The field of image segmentation has seen a significant increase in study in recent years, leading to the development of multi-level thresholding algorithms that have proven to be very successful compared to other approaches. Traditional statical techniques such as Otsu and Kapur are benchmark algorithms for image thresholding automation. The two techniques widely used, Otsu's and Kapur's entropy, are combined to create a hybrid fitness function to identify the ideal threshold values. In this study, we effectively reduce the computational time demonstrated by the high convergence curve while maintaining optimal outcomes by integrating the hybrid fitness function with multi-level thresholding using the Electric Eel Foraging Optimization (EEFO) approach to segment the uncovered region of masked face images. EEFO is a bio-inspired metaheuristic algorithm that simulates how electric EEL forages in nature. This algorithm achieved promising results in several optimization tasks, such as masked face segmentation. The proposed method is compared with ten cutting-edge algorithms focusing on recently developed metaheuristic techniques and outperforms them. Five metrics were used to evaluate the algorithm's performance: MSE, PSNR, SSIM, FSIM, and image quality index. The proposed method achieved superior results of 101.79, 26.83, 0.8058, 0.9339, and 0.9553 for average MSE, average PSNR, average SSIM, average FSIM, and average image quality index, respectively. Its superiority is verified by using the suggested approach on six benchmark images. The results demonstrate how effectively the proposed algorithm outperforms reliable metaheuristic approaches for solving masked face segmentation challenges.

由于越来越多地使用人脸面具，面具人脸分割任务变得更加困难。另一方面，前额、眉毛和眼睛区域通常是可见的，并能显示重要信息。人脸的这一暴露区域已被分割，并被信任地用于现实生活中的各种应用，如安全、医疗保健教育和智能城市项目。近年来，对图像分割领域的研究显著增加，导致了多级阈值算法的发展，与其他方法相比，这些算法已被证明是非常成功的。传统的静态技术（如 Otsu 和 Kapur）是图像阈值自动化的基准算法。大津熵和卡普尔熵这两种被广泛使用的技术被结合在一起，创造出一种混合拟合函数来识别理想的阈值。在本研究中，我们利用电鳗觅食优化（EEFO）方法将混合拟合函数与多级阈值相结合，对遮蔽人脸图像的未遮蔽区域进行分割，从而在保持最佳结果的同时，有效减少了高收敛曲线所显示的计算时间。EEFO 是一种受生物启发的元启发算法，它模拟了电鳗在自然界中的觅食方式。该算法在多项优化任务（如遮挡人脸分割）中取得了良好的效果。我们将所提出的方法与十种以最新开发的元启发式技术为重点的前沿算法进行了比较，结果发现该方法优于这些算法。该算法的性能评估采用了五个指标：MSE、PSNR、SSIM、FSIM 和图像质量指数。所提出的方法在平均 MSE、平均 PSNR、平均 SSIM、平均 FSIM 和平均图像质量指数方面分别取得了 101.79、26.83、0.8058、0.9339 和 0.9553 的优异成绩。通过在六幅基准图像上使用所建议的方法，验证了该算法的优越性。结果表明，在解决遮挡人脸分割难题方面，所建议的算法比可靠的元启发式方法更有效。

{"title":"Masked face image segmentation using a multilevel threshold with a hybrid fitness function","authors":"Nada AbdElFattah Ibrahim , Ehab R. Mohamed , Hanaa M. Hamza , Yousef S. Alsahafi , Khalid M. Hosny","doi":"10.1016/j.iswa.2024.200445","DOIUrl":"10.1016/j.iswa.2024.200445","url":null,"abstract":"<div><div>Masked face segmentation tasks have become significantly more difficult due to the increasing use of face masks. On the other hand, the forehead, eyebrows, and eye regions are usually visible and reveal vital information. This exposed area of the face has been segmented and trusted to be used in real life for various applications, such as security, healthcare education, and projects in smart cities. The field of image segmentation has seen a significant increase in study in recent years, leading to the development of multi-level thresholding algorithms that have proven to be very successful compared to other approaches. Traditional statical techniques such as Otsu and Kapur are benchmark algorithms for image thresholding automation. The two techniques widely used, Otsu's and Kapur's entropy, are combined to create a hybrid fitness function to identify the ideal threshold values. In this study, we effectively reduce the computational time demonstrated by the high convergence curve while maintaining optimal outcomes by integrating the hybrid fitness function with multi-level thresholding using the Electric Eel Foraging Optimization (EEFO) approach to segment the uncovered region of masked face images. EEFO is a bio-inspired metaheuristic algorithm that simulates how electric EEL forages in nature. This algorithm achieved promising results in several optimization tasks, such as masked face segmentation. The proposed method is compared with ten cutting-edge algorithms focusing on recently developed metaheuristic techniques and outperforms them. Five metrics were used to evaluate the algorithm's performance: MSE, PSNR, SSIM, FSIM, and image quality index. The proposed method achieved superior results of 101.79, 26.83, 0.8058, 0.9339, and 0.9553 for average MSE, average PSNR, average SSIM, average FSIM, and average image quality index, respectively. Its superiority is verified by using the suggested approach on six benchmark images. The results demonstrate how effectively the proposed algorithm outperforms reliable metaheuristic approaches for solving masked face segmentation challenges.</div></div>","PeriodicalId":100684,"journal":{"name":"Intelligent Systems with Applications","volume":"24 ","pages":"Article 200445"},"PeriodicalIF":0.0,"publicationDate":"2024-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142425941","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Deep learning for automated encrustation detection in sewer inspection 深度学习用于下水道检测中的自动结壳检测

Intelligent Systems with Applications

Pub Date : 2024-09-28 DOI: 10.1016/j.iswa.2024.200433

Wasiu Yusuf , Hafiz Alaka , Mubashir Ahmad , Wusu Godoyon , Saheed Ajayi , Luqman Olalekan Toriola-Coker , Abdullahi Ahmed

Rapid urbanization and population growth in recent decades have placed significant pressure on urban cities to rely heavily on underground infrastructure, such as sewers and tunnels, to maintain the provision of essential services. These sewers, typically having a limited lifespan of 50 to 100 years, are prone to various forms of defects. While prior research has primarily addressed common sewer defect like crack, root intrusion, and infiltration among others, the challenge of encrustation—the formation of hard deposits within sewer systems—has received less attention. This study presents a pioneering deep-learning approach to detect encrustation in sewers by leveraging survey videos from 14 different sewers in the United Kingdom. Our work marks the first effort to develop models specifically for detecting encrustation using deep learning techniques, as previous studies have focused on other types of deposits such as settled and attached deposits. By converting the videos into sequential image frames, we subjected them to thorough analysis and several image pre-processing techniques. Our contributions include the development and comparison of different classification models using backbone CNN networks such as AlexNet, VGG16, EfficientNet, and VGG19 to classify encrustation. Notably, this study provides the first metric-based comparison of these backbone networks to identify the most effective model for encrustation detection. The results demonstrate an impressive 96 % accuracy using the deep architecture of VGG19. Beyond accuracy, this research explores the impact of data augmentation and network dropout on reducing overfitting and enhancing model performance. Additionally, we analyze the time complexities associated with training models with and without data augmentation, providing valuable insights into the efficiency of our approach.

近几十年来，快速的城市化和人口增长给城市带来了巨大压力，使其不得不严重依赖下水道和隧道等地下基础设施来维持基本服务的提供。这些下水道的使用寿命通常只有 50 到 100 年，很容易出现各种形式的缺陷。以往的研究主要针对常见的下水道缺陷，如裂缝、根系侵入和渗透等，而对于结壳难题--下水道系统内坚硬沉积物的形成--却关注较少。本研究利用英国 14 个不同下水道的调查视频，提出了一种检测下水道结壳的开创性深度学习方法。我们的工作标志着利用深度学习技术开发专门用于检测结壳的模型的首次尝试，因为之前的研究主要针对其他类型的沉积物，如沉降和附着沉积物。通过将视频转换为连续图像帧，我们对其进行了全面分析，并采用了多种图像预处理技术。我们的贡献包括开发和比较了使用 AlexNet、VGG16、EfficientNet 和 VGG19 等骨干 CNN 网络对包壳进行分类的不同分类模型。值得注意的是，本研究首次对这些骨干网络进行了基于度量的比较，以确定最有效的结壳检测模型。结果表明，使用 VGG19 的深度架构，准确率达到了令人印象深刻的 96%。除了准确率，本研究还探讨了数据扩增和网络剔除对减少过拟合和提高模型性能的影响。此外，我们还分析了使用和不使用数据增强训练模型的时间复杂性，为了解我们方法的效率提供了宝贵的见解。

{"title":"Deep learning for automated encrustation detection in sewer inspection","authors":"Wasiu Yusuf , Hafiz Alaka , Mubashir Ahmad , Wusu Godoyon , Saheed Ajayi , Luqman Olalekan Toriola-Coker , Abdullahi Ahmed","doi":"10.1016/j.iswa.2024.200433","DOIUrl":"10.1016/j.iswa.2024.200433","url":null,"abstract":"<div><div>Rapid urbanization and population growth in recent decades have placed significant pressure on urban cities to rely heavily on underground infrastructure, such as sewers and tunnels, to maintain the provision of essential services. These sewers, typically having a limited lifespan of 50 to 100 years, are prone to various forms of defects. While prior research has primarily addressed common sewer defect like crack, root intrusion, and infiltration among others, the challenge of encrustation—the formation of hard deposits within sewer systems—has received less attention. This study presents a pioneering deep-learning approach to detect encrustation in sewers by leveraging survey videos from 14 different sewers in the United Kingdom. Our work marks the first effort to develop models specifically for detecting encrustation using deep learning techniques, as previous studies have focused on other types of deposits such as settled and attached deposits. By converting the videos into sequential image frames, we subjected them to thorough analysis and several image pre-processing techniques. Our contributions include the development and comparison of different classification models using backbone CNN networks such as AlexNet, VGG16, EfficientNet, and VGG19 to classify encrustation. Notably, this study provides the first metric-based comparison of these backbone networks to identify the most effective model for encrustation detection. The results demonstrate an impressive 96 % accuracy using the deep architecture of VGG19. Beyond accuracy, this research explores the impact of data augmentation and network dropout on reducing overfitting and enhancing model performance. Additionally, we analyze the time complexities associated with training models with and without data augmentation, providing valuable insights into the efficiency of our approach.</div></div>","PeriodicalId":100684,"journal":{"name":"Intelligent Systems with Applications","volume":"24 ","pages":"Article 200433"},"PeriodicalIF":0.0,"publicationDate":"2024-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142525833","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optimizing feature extraction and fusion for high-resolution defect detection in solar cells 优化特征提取和融合，实现太阳能电池的高分辨率缺陷检测

Intelligent Systems with Applications

Pub Date : 2024-09-26 DOI: 10.1016/j.iswa.2024.200443

Hoanh Nguyen, Tuan Anh Nguyen, Nguyen Duc Toan

In this paper, we propose a novel architecture for defect detection in electroluminescent images of polycrystalline silicon solar cells, addressing the challenges posed by subtle and dispersed defects. Our model, based on a modified Swin Transformer, incorporates key innovations that enhance feature extraction and fusion. We replace the conventional self-attention mechanism with a novel group self-attention mechanism, increasing the mAP50:5:95 score from 50.12 % to 52.98 % while reducing inference time from 74 ms to 62 ms. We also introduce a spatial displacement with shift convolution module, replacing the traditional Multi-Layer Perceptron, which further enhances the model's receptive field and improves precision and recall. Additionally, our fast multi-scale feature fusion mechanism effectively combines high-resolution details with high-level semantic features from different network layers, optimizing defect detection accuracy. Experimental results on the PVEL-AD dataset demonstrate that our model achieves the highest mAP50 score of 83.11 % and an F1-Score of 84.33 %, surpassing state-of-the-art models while maintaining a competitive inference time of 66.3 ms. These findings highlight the effectiveness of our innovations in improving defect detection accuracy and computational efficiency, making our model a robust solution for quality assurance in solar cell manufacturing.

在本文中，我们提出了一种用于多晶硅太阳能电池电致发光图像中缺陷检测的新型架构，以应对细微和分散缺陷带来的挑战。我们的模型以改进的斯温变换器为基础，融入了增强特征提取和融合的关键创新技术。我们用一种新颖的群组自我注意机制取代了传统的自我注意机制，将 mAP50:5:95 分数从 50.12% 提高到 52.98%，同时将推理时间从 74 毫秒缩短到 62 毫秒。我们还引入了带移位卷积的空间位移模块，取代了传统的多层感知器，从而进一步增强了模型的感受野，提高了精确度和召回率。此外，我们的快速多尺度特征融合机制有效地结合了高分辨率细节和来自不同网络层的高层次语义特征，优化了缺陷检测的准确性。在 PVEL-AD 数据集上的实验结果表明，我们的模型获得了 83.11 % 的最高 mAP50 分数和 84.33 % 的 F1 分数，超越了最先进的模型，同时保持了 66.3 毫秒的极具竞争力的推理时间。这些发现凸显了我们的创新在提高缺陷检测准确性和计算效率方面的有效性，使我们的模型成为太阳能电池制造质量保证的强大解决方案。

{"title":"Optimizing feature extraction and fusion for high-resolution defect detection in solar cells","authors":"Hoanh Nguyen, Tuan Anh Nguyen, Nguyen Duc Toan","doi":"10.1016/j.iswa.2024.200443","DOIUrl":"10.1016/j.iswa.2024.200443","url":null,"abstract":"<div><div>In this paper, we propose a novel architecture for defect detection in electroluminescent images of polycrystalline silicon solar cells, addressing the challenges posed by subtle and dispersed defects. Our model, based on a modified Swin Transformer, incorporates key innovations that enhance feature extraction and fusion. We replace the conventional self-attention mechanism with a novel group self-attention mechanism, increasing the mAP50:5:95 score from 50.12 % to 52.98 % while reducing inference time from 74 ms to 62 ms. We also introduce a spatial displacement with shift convolution module, replacing the traditional Multi-Layer Perceptron, which further enhances the model's receptive field and improves precision and recall. Additionally, our fast multi-scale feature fusion mechanism effectively combines high-resolution details with high-level semantic features from different network layers, optimizing defect detection accuracy. Experimental results on the PVEL-AD dataset demonstrate that our model achieves the highest mAP50 score of 83.11 % and an F1-Score of 84.33 %, surpassing state-of-the-art models while maintaining a competitive inference time of 66.3 ms. These findings highlight the effectiveness of our innovations in improving defect detection accuracy and computational efficiency, making our model a robust solution for quality assurance in solar cell manufacturing.</div></div>","PeriodicalId":100684,"journal":{"name":"Intelligent Systems with Applications","volume":"24 ","pages":"Article 200443"},"PeriodicalIF":0.0,"publicationDate":"2024-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142425938","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Hybrid intelligent algorithm aided energy consumption optimization in smart grid systems with edge computing 采用边缘计算的智能电网系统中的混合智能算法辅助能耗优化

Intelligent Systems with Applications

Pub Date : 2024-09-26 DOI: 10.1016/j.iswa.2024.200444

Shuangwei Li, Yang Xie, Mingming Shi, Xian Zheng, Yongling Lu

The rapid proliferation of smart grid systems necessitates efficient management of energy resources, particularly in the context of mobile edge computing (MEC) networks. This paper presents a novel approach to optimize the energy consumption in smart grid systems with the integration of edge computing, employing a hybrid intelligent algorithm (HIA) empowered by particle swarm optimization (PSO). The primary objective is to enhance the sustainability and operational efficiency of smart grid infrastructures by minimizing the energy consumption in the MEC networks. The proposed HIA utilizes PSO to dynamically allocate computational tasks and manage resources among edge devices based on real-time demand fluctuations. This adaptive approach aims to achieve the optimal load balancing and energy efficiency across the smart grid ecosystem. By leveraging the PSO’s ability to iteratively refine solutions and adapt to changing environmental conditions, the algorithm optimizes the energy consumption while maintaining requisite service levels and reliability. Simulation experiments and case studies validate the effectiveness of the proposed PSO-based HIA in reducing the energy consumption without compromising system other performances. The results demonstrate substantial improvements in the energy efficiency, illustrating the feasibility and benefits of employing intelligent algorithms tailored for edge computing environments within smart grid systems. This research contributes to advancing sustainable smart grid technologies by introducing a robust framework for energy optimization through hybrid intelligent algorithms.

智能电网系统的迅速普及要求对能源资源进行有效管理，尤其是在移动边缘计算（MEC）网络的背景下。本文采用粒子群优化（PSO）赋予的混合智能算法（HIA），提出了一种优化智能电网系统能源消耗与边缘计算整合的新方法。其主要目标是通过最大限度地降低 MEC 网络的能耗，提高智能电网基础设施的可持续性和运行效率。拟议的 HIA 利用 PSO 根据实时需求波动在边缘设备之间动态分配计算任务和管理资源。这种自适应方法旨在实现整个智能电网生态系统的最佳负载平衡和能效。通过利用 PSO 的迭代改进解决方案和适应不断变化的环境条件的能力，该算法在保持必要的服务水平和可靠性的同时优化了能源消耗。仿真实验和案例研究验证了所提出的基于 PSO 的 HIA 在降低能耗而不影响系统其他性能方面的有效性。结果表明，能效有了显著提高，说明了在智能电网系统中采用专为边缘计算环境定制的智能算法的可行性和益处。这项研究通过混合智能算法引入了一个稳健的能源优化框架，为推动可持续智能电网技术的发展做出了贡献。

{"title":"Hybrid intelligent algorithm aided energy consumption optimization in smart grid systems with edge computing","authors":"Shuangwei Li, Yang Xie, Mingming Shi, Xian Zheng, Yongling Lu","doi":"10.1016/j.iswa.2024.200444","DOIUrl":"10.1016/j.iswa.2024.200444","url":null,"abstract":"<div><div>The rapid proliferation of smart grid systems necessitates efficient management of energy resources, particularly in the context of mobile edge computing (MEC) networks. This paper presents a novel approach to optimize the energy consumption in smart grid systems with the integration of edge computing, employing a hybrid intelligent algorithm (HIA) empowered by particle swarm optimization (PSO). The primary objective is to enhance the sustainability and operational efficiency of smart grid infrastructures by minimizing the energy consumption in the MEC networks. The proposed HIA utilizes PSO to dynamically allocate computational tasks and manage resources among edge devices based on real-time demand fluctuations. This adaptive approach aims to achieve the optimal load balancing and energy efficiency across the smart grid ecosystem. By leveraging the PSO’s ability to iteratively refine solutions and adapt to changing environmental conditions, the algorithm optimizes the energy consumption while maintaining requisite service levels and reliability. Simulation experiments and case studies validate the effectiveness of the proposed PSO-based HIA in reducing the energy consumption without compromising system other performances. The results demonstrate substantial improvements in the energy efficiency, illustrating the feasibility and benefits of employing intelligent algorithms tailored for edge computing environments within smart grid systems. This research contributes to advancing sustainable smart grid technologies by introducing a robust framework for energy optimization through hybrid intelligent algorithms.</div></div>","PeriodicalId":100684,"journal":{"name":"Intelligent Systems with Applications","volume":"24 ","pages":"Article 200444"},"PeriodicalIF":0.0,"publicationDate":"2024-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142425939","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Power quality disturbances categorization using Identity Feature Vector and Extreme Learning Machine 利用身份特征向量和极限学习机进行电能质量干扰分类

Intelligent Systems with Applications

Pub Date : 2024-09-24 DOI: 10.1016/j.iswa.2024.200446

Shen Wei , Du Wenjuan , Chen Xia

Power quality disturbances are variations or anomalies in the voltage, current, or frequency of electrical power that can affect the proper operation of electrical equipment. These disturbances are usually classified into different categories based on their attributes and effects. This article presents an intelligent technique based on an Identity Feature Vector and an Extreme Learning Machine (ELM). This study first derives a constant length vector for each disturbance signal. A wavelet transform is applied to derive attributes from the input disturbance signal, and the identity vector is formed using the approximation coefficients. After the required normalization procedures, the normalized identity vector is classified using an ELM. To assess the productivity of the suggested approach, 12 types of disturbances, single and combined, are generated, and the system's efficiency is studied. The results indicate that ten out of 12 combinations, including Harmonic, Sag, and Flicker, were detected with 100 % accuracy. Additionally, the combination "Harmonic + Swell" exhibited the lowest accuracy, identified with 98 % accuracy. The total average accuracy of this method is 99.75 %. The outcomes demonstrate the highly favorable performance of this approach. This study evaluated the analyzed algorithm under noisy conditions with three different noise levels: 30 dB, 40 dB, and 50 dB, respectively. The average prediction accuracy for these three noise levels is 99.16 %, 99.25 %, and 98.91 %. The outcomes demonstrate that the evaluated algorithm accurately detects power quality disturbances across various noisy conditions.

电能质量干扰是指电压、电流或频率的变化或异常，会影响电气设备的正常运行。这些干扰通常根据其属性和影响分为不同的类别。本文介绍了一种基于身份特征向量和极限学习机（ELM）的智能技术。这项研究首先为每个干扰信号推导出一个恒定长度的向量。应用小波变换从输入干扰信号中提取属性，并利用近似系数形成特征向量。经过所需的归一化程序后，使用 ELM 对归一化特征向量进行分类。为了评估所建议方法的效率，生成了 12 种单一和组合干扰，并对系统的效率进行了研究。结果表明，在 12 种组合中，包括谐波、矢量和闪烁在内的 10 种组合的检测准确率达到了 100%。此外，"谐波 + 闪烁 "组合的准确率最低，仅为 98%。该方法的总平均准确率为 99.75%。这些结果表明，这种方法的性能非常出色。本研究在三种不同噪声水平（分别为 30 dB、40 dB 和 50 dB）的噪声条件下对所分析的算法进行了评估。这三种噪声水平的平均预测准确率分别为 99.16 %、99.25 % 和 98.91 %。结果表明，所评估的算法能在各种噪声条件下准确检测到电能质量干扰。

{"title":"Power quality disturbances categorization using Identity Feature Vector and Extreme Learning Machine","authors":"Shen Wei , Du Wenjuan , Chen Xia","doi":"10.1016/j.iswa.2024.200446","DOIUrl":"10.1016/j.iswa.2024.200446","url":null,"abstract":"<div><div>Power quality disturbances are variations or anomalies in the voltage, current, or frequency of electrical power that can affect the proper operation of electrical equipment. These disturbances are usually classified into different categories based on their attributes and effects. This article presents an intelligent technique based on an Identity Feature Vector and an Extreme Learning Machine (ELM). This study first derives a constant length vector for each disturbance signal. A wavelet transform is applied to derive attributes from the input disturbance signal, and the identity vector is formed using the approximation coefficients. After the required normalization procedures, the normalized identity vector is classified using an ELM. To assess the productivity of the suggested approach, 12 types of disturbances, single and combined, are generated, and the system's efficiency is studied. The results indicate that ten out of 12 combinations, including Harmonic, Sag, and Flicker, were detected with 100 % accuracy. Additionally, the combination \"Harmonic + Swell\" exhibited the lowest accuracy, identified with 98 % accuracy. The total average accuracy of this method is 99.75 %. The outcomes demonstrate the highly favorable performance of this approach. This study evaluated the analyzed algorithm under noisy conditions with three different noise levels: 30 dB, 40 dB, and 50 dB, respectively. The average prediction accuracy for these three noise levels is 99.16 %, 99.25 %, and 98.91 %. The outcomes demonstrate that the evaluated algorithm accurately detects power quality disturbances across various noisy conditions.</div></div>","PeriodicalId":100684,"journal":{"name":"Intelligent Systems with Applications","volume":"24 ","pages":"Article 200446"},"PeriodicalIF":0.0,"publicationDate":"2024-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142525803","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Enhanced intrusion detection model based on principal component analysis and variable ensemble machine learning algorithm 基于主成分分析和变量集合机器学习算法的增强型入侵检测模型

Intelligent Systems with Applications

Pub Date : 2024-09-21 DOI: 10.1016/j.iswa.2024.200442

Ayuba John , Ismail Fauzi Bin Isnin , Syed Hamid Hussain Madni , Farkhana Binti Muchtar

The intrusion detection system (IDS) model, which can identify the presence of intruders in the network and take some predefined action for safe data transit across the network, is advantageous in achieving security in both simple and advanced network systems. Several IDS models have various security problems, such as low detection accuracy and high false alarms, which can be caused by the network traffic dataset's excessive dimensionality and class imbalance in the creation of IDS models. Principal Component Analysis (PCA) has proven to be a helpful feature selection technique for dimensionality reduction. As a result, because it is a linear transformation, it has challenges capturing non-linear relationships between feature properties in the network traffic datasets. This paper proposes a variable ensemble machine learning method to solve the problem and achieve a low variance model with high accuracy and low false alarm. First, PCA is combined with the AdaBoost ensemble machine learning algorithm, which acts as stagewise additive modelling to compensate for PCA's deficiency in feature selection in network traffic by minimizing the exponential loss function. Secondly, PCA is used for feature selection, and a LogitBoost classifier algorithm can be used for multiclass classification and acts as an additive tree regression to compensate for the PCA's weakness by minimizing the Logistic Loss to provide an optimal classifier output. Finally, the low variance ability of RandomForest, which employs the bagging approach, is applied to eliminate overfittings. The experiments of the IDS model developed from the proposed methods were evaluated on the WSN-DS, NSL-KDD, and UNSW-N15 datasets. The performance of the methods, PCA with AdaBoost, on the WSN-DS dataset has an accuracy score of 92.3 %, an 89.0 % accuracy score on the NSL-KDD dataset, and a 67.9 % accuracy score on UNSW-N15, which is the least accurate score. PCA and RandomForest surpassed them by scoring 100 % accuracy on all three datasets. PCA and Bagging have an accuracy score of 99.8 % on the WSN-DS dataset, 100 % on the NSL-KDD dataset, and 93.4 % on the UNSW-N15 dataset. In comparison, PCA and LogitBoost have an accuracy score of 98.9 % on the WSN-DS dataset, 100 % on the NSL-KDD dataset, and 88.7 % on the UNSW-N15 dataset.

入侵检测系统（IDS）模型可以识别网络中是否存在入侵者，并采取一些预定义的措施以确保数据在网络中的安全传输，它在实现简单和高级网络系统的安全性方面都具有优势。一些 IDS 模型存在各种安全问题，如检测准确率低和误报率高，这可能是由于创建 IDS 模型时网络流量数据集的维度过大和类不平衡造成的。事实证明，主成分分析（PCA）是一种有助于降维的特征选择技术。但由于它是一种线性变换，因此在捕捉网络流量数据集中特征属性之间的非线性关系方面存在挑战。本文提出了一种变量集合机器学习方法来解决这一问题，并实现了高精度、低误报的低方差模型。首先，将 PCA 与 AdaBoost 集合机器学习算法相结合，通过最小化指数损失函数，发挥阶段性加法建模的作用，弥补 PCA 在网络流量特征选择方面的不足。其次，PCA 用于特征选择，LogitBoost 分类器算法可用于多类分类，作为加法树回归，通过最小化 Logistic 损失来弥补 PCA 的不足，从而提供最佳分类器输出。最后，随机森林（RandomForest）的低方差能力采用了袋集方法，以消除过拟合。在 WSN-DS、NSL-KDD 和 UNSW-N15 数据集上对根据所提方法开发的 IDS 模型进行了实验评估。PCA 和 AdaBoost 方法在 WSN-DS 数据集上的准确率为 92.3%，在 NSL-KDD 数据集上的准确率为 89.0%，在 UNSW-N15 数据集上的准确率为 67.9%，是准确率最低的数据集。PCA 和 RandomForest 在这三个数据集上的准确率都达到了 100%，超过了它们。PCA 和 Bagging 在 WSN-DS 数据集上的准确率为 99.8%，在 NSL-KDD 数据集上为 100%，在 UNSW-N15 数据集上为 93.4%。相比之下，PCA 和 LogitBoost 在 WSN-DS 数据集上的准确率为 98.9%，在 NSL-KDD 数据集上为 100%，在 UNSW-N15 数据集上为 88.7%。

{"title":"Enhanced intrusion detection model based on principal component analysis and variable ensemble machine learning algorithm","authors":"Ayuba John , Ismail Fauzi Bin Isnin , Syed Hamid Hussain Madni , Farkhana Binti Muchtar","doi":"10.1016/j.iswa.2024.200442","DOIUrl":"10.1016/j.iswa.2024.200442","url":null,"abstract":"<div><div>The intrusion detection system (IDS) model, which can identify the presence of intruders in the network and take some predefined action for safe data transit across the network, is advantageous in achieving security in both simple and advanced network systems. Several IDS models have various security problems, such as low detection accuracy and high false alarms, which can be caused by the network traffic dataset's excessive dimensionality and class imbalance in the creation of IDS models. Principal Component Analysis (PCA) has proven to be a helpful feature selection technique for dimensionality reduction. As a result, because it is a linear transformation, it has challenges capturing non-linear relationships between feature properties in the network traffic datasets. This paper proposes a variable ensemble machine learning method to solve the problem and achieve a low variance model with high accuracy and low false alarm. First, PCA is combined with the AdaBoost ensemble machine learning algorithm, which acts as stagewise additive modelling to compensate for PCA's deficiency in feature selection in network traffic by minimizing the exponential loss function. Secondly, PCA is used for feature selection, and a LogitBoost classifier algorithm can be used for multiclass classification and acts as an additive tree regression to compensate for the PCA's weakness by minimizing the Logistic Loss to provide an optimal classifier output. Finally, the low variance ability of RandomForest, which employs the bagging approach, is applied to eliminate overfittings. The experiments of the IDS model developed from the proposed methods were evaluated on the WSN-DS, NSL-KDD, and UNSW-N15 datasets. The performance of the methods, PCA with AdaBoost, on the WSN-DS dataset has an accuracy score of 92.3 %, an 89.0 % accuracy score on the NSL-KDD dataset, and a 67.9 % accuracy score on UNSW-N15, which is the least accurate score. PCA and RandomForest surpassed them by scoring 100 % accuracy on all three datasets. PCA and Bagging have an accuracy score of 99.8 % on the WSN-DS dataset, 100 % on the NSL-KDD dataset, and 93.4 % on the UNSW-N15 dataset. In comparison, PCA and LogitBoost have an accuracy score of 98.9 % on the WSN-DS dataset, 100 % on the NSL-KDD dataset, and 88.7 % on the UNSW-N15 dataset.</div></div>","PeriodicalId":100684,"journal":{"name":"Intelligent Systems with Applications","volume":"24 ","pages":"Article 200442"},"PeriodicalIF":0.0,"publicationDate":"2024-09-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142324052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Evolution of machine learning applications in medical and healthcare analytics research: A bibliometric analysis 医疗和保健分析研究中机器学习应用的演变：文献计量分析

Intelligent Systems with Applications

Pub Date : 2024-09-19 DOI: 10.1016/j.iswa.2024.200441

Samuel-Soma M. Ajibade , Gloria Nnadwa Alhassan , Abdelhamid Zaidi , Olukayode Ayodele Oki , Joseph Bamidele Awotunde , Emeka Ogbuju , Kayode A. Akintoye

This bibliometric research explores the global evolution of machine learning applications in medical and healthcare research for 3 decades (1994 to 2023). The study applies data mining techniques to a comprehensive dataset of published articles related to machine learning applications in the medical and healthcare sectors. The data extraction process includes the retrieval of relevant information from the source sources such as journals, books, and conference proceedings. An analysis of the extracted data is then conducted to identify the trends in the machine learning applications in medical and healthcare research. The Results revealed the publications published and indexed in the Scopus and PubMed database over the last 30 years. Bibliometric Analysis revealed that funding played a more significant role in publication productivity compared to collaboration (co-authorships), particularly at the country level. Hotspots analysis revealed three core research themes on MLHC research hence demonstrating the importance of machine learning applications to medical and healthcare research. Further, the study showed that the MLHC research landscape has largely focused on ML applications to tackle various issues ranging from chronic medical challenges (e.g., cardiological diseases) to patient data security. The findings of this research may be useful to policy makers and practitioners in the medical and healthcare sectors and to global research endeavours in the field. Future studies could include addressing issues such as growing ethical considerations, integration, and practical applications in wearable technology, IoT, and smart healthcare systems.

这项文献计量学研究探讨了机器学习在医疗和保健研究领域的应用在三十年间（1994 年至 2023 年）的全球演变情况。该研究将数据挖掘技术应用于一个全面的数据集，该数据集包含与医疗和保健领域机器学习应用相关的已发表文章。数据提取过程包括从期刊、书籍和会议论文集等来源中检索相关信息。然后对提取的数据进行分析，以确定医疗和保健研究中机器学习应用的趋势。结果显示了过去 30 年中在 Scopus 和 PubMed 数据库中发表和索引的出版物。文献计量分析表明，与合作（合著）相比，资金在论文发表率中发挥着更重要的作用，尤其是在国家层面。热点分析揭示了 MLHC 研究的三个核心研究主题，从而证明了机器学习应用在医疗和保健研究中的重要性。此外，研究还表明，多语言医疗保健的研究领域主要集中在机器学习应用方面，以解决从慢性医疗挑战（如心脏病）到患者数据安全等各种问题。这项研究的结果可能对医疗和保健领域的政策制定者和从业人员以及该领域的全球研究工作有所帮助。未来的研究可能包括解决可穿戴技术、物联网和智能医疗系统中日益增长的伦理考虑、整合和实际应用等问题。

{"title":"Evolution of machine learning applications in medical and healthcare analytics research: A bibliometric analysis","authors":"Samuel-Soma M. Ajibade , Gloria Nnadwa Alhassan , Abdelhamid Zaidi , Olukayode Ayodele Oki , Joseph Bamidele Awotunde , Emeka Ogbuju , Kayode A. Akintoye","doi":"10.1016/j.iswa.2024.200441","DOIUrl":"10.1016/j.iswa.2024.200441","url":null,"abstract":"<div><div>This bibliometric research explores the global evolution of machine learning applications in medical and healthcare research for 3 decades (1994 to 2023). The study applies data mining techniques to a comprehensive dataset of published articles related to machine learning applications in the medical and healthcare sectors. The data extraction process includes the retrieval of relevant information from the source sources such as journals, books, and conference proceedings. An analysis of the extracted data is then conducted to identify the trends in the machine learning applications in medical and healthcare research. The Results revealed the publications published and indexed in the Scopus and PubMed database over the last 30 years. Bibliometric Analysis revealed that funding played a more significant role in publication productivity compared to collaboration (co-authorships), particularly at the country level. Hotspots analysis revealed three core research themes on MLHC research hence demonstrating the importance of machine learning applications to medical and healthcare research. Further, the study showed that the MLHC research landscape has largely focused on ML applications to tackle various issues ranging from chronic medical challenges (e.g., cardiological diseases) to patient data security. The findings of this research may be useful to policy makers and practitioners in the medical and healthcare sectors and to global research endeavours in the field. Future studies could include addressing issues such as growing ethical considerations, integration, and practical applications in wearable technology, IoT, and smart healthcare systems.</div></div>","PeriodicalId":100684,"journal":{"name":"Intelligent Systems with Applications","volume":"24 ","pages":"Article 200441"},"PeriodicalIF":0.0,"publicationDate":"2024-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142326554","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optimization of inventory management through computer vision and machine learning technologies 通过计算机视觉和机器学习技术优化库存管理

Intelligent Systems with Applications

Pub Date : 2024-09-19 DOI: 10.1016/j.iswa.2024.200438

William Villegas-Ch , Alexandra Maldonado Navarro , Santiago Sanchez-Viteri

This study presents implementing and evaluating a computer vision platform to optimize warehouse inventory management. Integrating machine learning and computer vision technologies, this solution addresses critical challenges in inventory accuracy and operational efficiency, overcoming the limitations of traditional methods and pre-existing automated systems. The platform uses convolutional neural networks and open-source libraries such as TensorFlow and PyTorch to recognize and accurately classify products from images captured in real time. Practical implementation in a natural warehouse environment allowed the proposed platform to be compared with traditional systems, highlighting significant improvements, such as a 45% reduction in the time required for inventory counting and a 9% increase in inventory accuracy. Despite facing challenges such as staff resistance to change and technical limitations on image quality, these difficulties were overcome through effective change management strategies and algorithm improvements. The findings of this study identify the potential for computer vision technology to transform warehouse operations, offering a practical and adaptable solution for inventory management.

本研究介绍了计算机视觉平台的实施和评估情况，以优化仓库库存管理。该解决方案整合了机器学习和计算机视觉技术，克服了传统方法和现有自动化系统的局限性，解决了库存准确性和运营效率方面的关键挑战。该平台使用卷积神经网络以及 TensorFlow 和 PyTorch 等开源库，从实时捕获的图像中识别产品并对其进行准确分类。通过在自然仓库环境中的实际应用，可以将拟议的平台与传统系统进行比较，结果发现该平台有了显著改善，例如库存清点所需时间减少了 45%，库存准确率提高了 9%。尽管面临员工抵制变革和图像质量技术限制等挑战，但通过有效的变革管理策略和算法改进，这些困难都被克服了。这项研究的结果确定了计算机视觉技术改变仓库运作的潜力，为库存管理提供了一个实用且适应性强的解决方案。

{"title":"Optimization of inventory management through computer vision and machine learning technologies","authors":"William Villegas-Ch , Alexandra Maldonado Navarro , Santiago Sanchez-Viteri","doi":"10.1016/j.iswa.2024.200438","DOIUrl":"10.1016/j.iswa.2024.200438","url":null,"abstract":"<div><p>This study presents implementing and evaluating a computer vision platform to optimize warehouse inventory management. Integrating machine learning and computer vision technologies, this solution addresses critical challenges in inventory accuracy and operational efficiency, overcoming the limitations of traditional methods and pre-existing automated systems. The platform uses convolutional neural networks and open-source libraries such as TensorFlow and PyTorch to recognize and accurately classify products from images captured in real time. Practical implementation in a natural warehouse environment allowed the proposed platform to be compared with traditional systems, highlighting significant improvements, such as a 45% reduction in the time required for inventory counting and a 9% increase in inventory accuracy. Despite facing challenges such as staff resistance to change and technical limitations on image quality, these difficulties were overcome through effective change management strategies and algorithm improvements. The findings of this study identify the potential for computer vision technology to transform warehouse operations, offering a practical and adaptable solution for inventory management.</p></div>","PeriodicalId":100684,"journal":{"name":"Intelligent Systems with Applications","volume":"24 ","pages":"Article 200438"},"PeriodicalIF":0.0,"publicationDate":"2024-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2667305324001121/pdfft?md5=fddb74ba205cf2dbd45a73920fc45d01&pid=1-s2.0-S2667305324001121-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142272161","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

DeepInvesting: Stock market predictions with a sequence-oriented BiLSTM stacked model – A dataset case study of AMZN 深度投资：利用面向序列的 BiLSTM 叠加模型预测股市--AMZN 数据集案例研究

Intelligent Systems with Applications

Pub Date : 2024-09-19 DOI: 10.1016/j.iswa.2024.200439

Ashkan Safari, Mohammad Ali Badamchizadeh

Intelligent forecasters are now being considered in the stock market, providing essential insights and strategic guidance to investors and traders by presenting analytical tools and predictive models, thus enabling informed decision-making and mitigating financial risks in this dynamic market. The importance of intelligent analyzers in stock trading routines is considered in this work, where DeepInvesting, a multimodal deep learning model tailored for stock price prediction, is introduced. Employing a Sequence-Oriented, Long-Term Dependent (SoLTD) architecture featuring Bidirectional Long Short-Term Memory (BiLSTM) networks, DeepInvesting is applied to essential features of the Amazon Corp. (AMZN) market dataset, gathered from Yahoo Finance, including Closing, Opening, High, Low, Volume, and Adj Close prices. Exceptional performance in forecasting Closing, Opening, High, Low, and Adj Close prices is demonstrated, with minimal Mean Absolute Percentage Error (MAPE) and Root Mean Squared Percentage Error (RMSPE) scores, coupled with high R-squared (R²) values, manifesting a robust fit to the data, as well as computational complexity, and Rates Per Second (RPS) metrics in comparison to other models of KNN, LSTM, RNN, CNN, and ANN. Finally, challenges in the accurate prediction of trading volumes are identified, highlighting an area for future enhancement.

目前，股票市场正在考虑采用智能预测器，通过提供分析工具和预测模型，为投资者和交易者提供重要的见解和战略指导，从而在这个动态市场中做出明智决策并降低金融风险。本研究考虑了智能分析仪在股票交易程序中的重要性，介绍了专为股票价格预测定制的多模态深度学习模型 DeepInvesting。DeepInvesting 采用以双向长短期记忆（BiLSTM）网络为特色的面向序列、长期依赖（SoLTD）架构，应用于从雅虎财经收集的亚马逊公司（AMZN）市场数据集的基本特征，包括收盘价、开盘价、最高价、最低价、成交量和 Adj Close 价格。与 KNN、LSTM、RNN、CNN 和 ANN 等其他模型相比，该方法在预测收盘价、开盘价、最高价、最低价和 Adj Close 价格方面表现出色，平均绝对误差 (MAPE) 和均方根误差 (RMSPE) 分数最小，R 平方 (R2) 值高，与数据的拟合度以及计算复杂度和每秒速率 (RPS) 指标都很高。最后，还指出了在准确预测交易量方面存在的挑战，并强调了未来需要改进的领域。

{"title":"DeepInvesting: Stock market predictions with a sequence-oriented BiLSTM stacked model – A dataset case study of AMZN","authors":"Ashkan Safari, Mohammad Ali Badamchizadeh","doi":"10.1016/j.iswa.2024.200439","DOIUrl":"10.1016/j.iswa.2024.200439","url":null,"abstract":"<div><div>Intelligent forecasters are now being considered in the stock market, providing essential insights and strategic guidance to investors and traders by presenting analytical tools and predictive models, thus enabling informed decision-making and mitigating financial risks in this dynamic market. The importance of intelligent analyzers in stock trading routines is considered in this work, where DeepInvesting, a multimodal deep learning model tailored for stock price prediction, is introduced. Employing a Sequence-Oriented, Long-Term Dependent (SoLTD) architecture featuring Bidirectional Long Short-Term Memory (BiLSTM) networks, DeepInvesting is applied to essential features of the Amazon Corp. (AMZN) market dataset, gathered from Yahoo Finance, including Closing, Opening, High, Low, Volume, and Adj Close prices. Exceptional performance in forecasting Closing, Opening, High, Low, and Adj Close prices is demonstrated, with minimal Mean Absolute Percentage Error (MAPE) and Root Mean Squared Percentage Error (RMSPE) scores, coupled with high R-squared (R<sup>2</sup>) values, manifesting a robust fit to the data, as well as computational complexity, and Rates Per Second (RPS) metrics in comparison to other models of KNN, LSTM, RNN, CNN, and ANN. Finally, challenges in the accurate prediction of trading volumes are identified, highlighting an area for future enhancement.</div></div>","PeriodicalId":100684,"journal":{"name":"Intelligent Systems with Applications","volume":"24 ","pages":"Article 200439"},"PeriodicalIF":0.0,"publicationDate":"2024-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2667305324001133/pdfft?md5=7fc986df27640d36742f23b85b7b526b&pid=1-s2.0-S2667305324001133-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142315171","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0