Pub Date : 2022-12-16DOI: 10.1109/ICCWAMTIP56608.2022.10016612
Huang Chao
Image anomaly detection and localization not only need to provide image-level anomaly judgment but also need to locate pixel-level anomaly areas. This paper proposes a model named pixelAD, which builds an end-to-end network through pixel- level feature clustering learning to solve this problem. The normal prototype is obtained during training by clustering the normal pixel-level features. We generate pixel-level cluster labels of normal samples according to the prototypes, which guide the model to update parameters by calculating the assignment loss. For inference, pixelAD directly outputs the pixel-level anomaly score end-to-end. The experimental results of the real industrial dataset MVTecAD show that PixelAD has an excellent performance in anomaly detection and anomaly localization.
{"title":"Pixel-Level Feature Clustering Learning for Image Anomaly Detection and Localization","authors":"Huang Chao","doi":"10.1109/ICCWAMTIP56608.2022.10016612","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016612","url":null,"abstract":"Image anomaly detection and localization not only need to provide image-level anomaly judgment but also need to locate pixel-level anomaly areas. This paper proposes a model named pixelAD, which builds an end-to-end network through pixel- level feature clustering learning to solve this problem. The normal prototype is obtained during training by clustering the normal pixel-level features. We generate pixel-level cluster labels of normal samples according to the prototypes, which guide the model to update parameters by calculating the assignment loss. For inference, pixelAD directly outputs the pixel-level anomaly score end-to-end. The experimental results of the real industrial dataset MVTecAD show that PixelAD has an excellent performance in anomaly detection and anomaly localization.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132336032","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-16DOI: 10.1109/ICCWAMTIP56608.2022.10016603
Xiang Li, Laixiang Qiu, Yujun Yang, Wang Zhou
With the rapid development of the Internet, the amount of global data shows an explosive growth, and the phenomenon of information overload has become more and more serious. How to obtain the information that users really care about has become one of the problems that people need to overcome. In this background, many recommendation algorithms have been widely used in all walks of life. This paper combs the background and future development trend of several recommendation algorithms.
{"title":"Recommender Algorithms: From Matrix Factorization to Neural Network","authors":"Xiang Li, Laixiang Qiu, Yujun Yang, Wang Zhou","doi":"10.1109/ICCWAMTIP56608.2022.10016603","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016603","url":null,"abstract":"With the rapid development of the Internet, the amount of global data shows an explosive growth, and the phenomenon of information overload has become more and more serious. How to obtain the information that users really care about has become one of the problems that people need to overcome. In this background, many recommendation algorithms have been widely used in all walks of life. This paper combs the background and future development trend of several recommendation algorithms.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"148 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127243273","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-16DOI: 10.1109/ICCWAMTIP56608.2022.10016482
Zeng Boheng
Vision Transformers (ViTs) have recently made great improvements in numerous vision tasks. To safely deploy in real life, it is crucial to investigate the vulnerability of ViTs. Model augmentation is an effective way to improve cross-model transferability. However, recent works mainly focus on deep neural networks (DNNs) but present a low transferability to ViTs due to the lack of usage of their properties. Inspired by the fact that ViTs are insensitive to spatial structure, we conduct experiments and demonstrated that ViTs have similar cross entropy for disrupted images, which we called perturbation invariance. Therefore, we propose our perturbation invariance method to improve transferability. Specifically, we craft transformed images by randomly shuffling the input image patches, and average the gradients of these transformed images each iteration. Besides, we also add Gaussian noise in the iterative process to further boost attack. Extensive experiments prove the effectiveness of our method. In particular, our method obtains a 79.5% attack success rate on average when against four types of ViTs, which outperforms other state-ofthe-art methods by 3.8%.
{"title":"Enhancing the Adversarial Transferability of Vision Transformers Through Perturbation Invariance","authors":"Zeng Boheng","doi":"10.1109/ICCWAMTIP56608.2022.10016482","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016482","url":null,"abstract":"Vision Transformers (ViTs) have recently made great improvements in numerous vision tasks. To safely deploy in real life, it is crucial to investigate the vulnerability of ViTs. Model augmentation is an effective way to improve cross-model transferability. However, recent works mainly focus on deep neural networks (DNNs) but present a low transferability to ViTs due to the lack of usage of their properties. Inspired by the fact that ViTs are insensitive to spatial structure, we conduct experiments and demonstrated that ViTs have similar cross entropy for disrupted images, which we called perturbation invariance. Therefore, we propose our perturbation invariance method to improve transferability. Specifically, we craft transformed images by randomly shuffling the input image patches, and average the gradients of these transformed images each iteration. Besides, we also add Gaussian noise in the iterative process to further boost attack. Extensive experiments prove the effectiveness of our method. In particular, our method obtains a 79.5% attack success rate on average when against four types of ViTs, which outperforms other state-ofthe-art methods by 3.8%.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127807132","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-16DOI: 10.1109/ICCWAMTIP56608.2022.10016546
Wang Han
Graph embedding has attracted increasing attention in the past few years. Graph neural network is the most popular type of algorithm to provide high-quality graph embedding. It brings stunning success but causes an important and easy-to-be-ignored problem, i.e., vulnerability, which makes the embedding quality dramatically degrade and affects the performance of downstream tasks like node classification. In this paper, we propose an algorithm for robust graph embedding via self-supervised graph denoising (SSGD). The key idea is to learn normal patterns about how a graph is organized and apply the patterns to reorganize the structure and remove noisy edges in the graph. Since the vulnerability is mainly caused by noisy edges, graph neural networks are supposed to work well on denoised graphs. In experiments, we introduce 6 state-of-the-art algorithms and 3 real-world datasets to demonstrate the superiority of our algorithm.
{"title":"Robust Graph Embedding via Self-Supervised Graph Denoising","authors":"Wang Han","doi":"10.1109/ICCWAMTIP56608.2022.10016546","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016546","url":null,"abstract":"Graph embedding has attracted increasing attention in the past few years. Graph neural network is the most popular type of algorithm to provide high-quality graph embedding. It brings stunning success but causes an important and easy-to-be-ignored problem, i.e., vulnerability, which makes the embedding quality dramatically degrade and affects the performance of downstream tasks like node classification. In this paper, we propose an algorithm for robust graph embedding via self-supervised graph denoising (SSGD). The key idea is to learn normal patterns about how a graph is organized and apply the patterns to reorganize the structure and remove noisy edges in the graph. Since the vulnerability is mainly caused by noisy edges, graph neural networks are supposed to work well on denoised graphs. In experiments, we introduce 6 state-of-the-art algorithms and 3 real-world datasets to demonstrate the superiority of our algorithm.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128485572","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-16DOI: 10.1109/ICCWAMTIP56608.2022.10016500
Li Xianli, Pan Wei, Wang Yisheng, Li Ming, Liu Guosong
The Internet of Things is widely used in industry, agriculture, health, urban management and other fields, the sensing terminal is an important part of IoT system, the security of sensing terminal directly affects the whole security of IoT system. This paper proposes the corresponding security measures of the sensing terminal of IoT system, such as physical security, access security, communication security, equipment security, data security, furthermore, we have verified it with experiments. It is of great significance for the selection, deployment, operation and maintenance the sensing terminal of IoT system, and has important application value for designing and producing the sensing terminal of IoT system.
{"title":"Research on Security Technology of Sensing Terminal of Internet of Things","authors":"Li Xianli, Pan Wei, Wang Yisheng, Li Ming, Liu Guosong","doi":"10.1109/ICCWAMTIP56608.2022.10016500","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016500","url":null,"abstract":"The Internet of Things is widely used in industry, agriculture, health, urban management and other fields, the sensing terminal is an important part of IoT system, the security of sensing terminal directly affects the whole security of IoT system. This paper proposes the corresponding security measures of the sensing terminal of IoT system, such as physical security, access security, communication security, equipment security, data security, furthermore, we have verified it with experiments. It is of great significance for the selection, deployment, operation and maintenance the sensing terminal of IoT system, and has important application value for designing and producing the sensing terminal of IoT system.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131678434","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-16DOI: 10.1109/ICCWAMTIP56608.2022.10016492
Jiquan Li, Laixiang Qiu, Yujun Yang, Wang Zhou
Breast cancer is one of the highest incidence in women's cancer, The pathological diagnosis of breast cancer can be used to evaluate the invasion of tumors and provide important information for accurate diagnosis and treatment. Statistics the number of mitosis cells in breast cancer is one of the important indicators of breast cancer division. In this paper, we summarized the current mainstream methods of mitosis cells detection. These methods are mainly implemented based on deep learning, and discussing the results of some of the methods, comparison and evaluation. At last, through the review of the research methods in this field, the existing breast cancer research methods have been summarized, and the future developments are prospected.
{"title":"Deep Learning Techniques for Breast Cancer Mitotic Cell Detection","authors":"Jiquan Li, Laixiang Qiu, Yujun Yang, Wang Zhou","doi":"10.1109/ICCWAMTIP56608.2022.10016492","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016492","url":null,"abstract":"Breast cancer is one of the highest incidence in women's cancer, The pathological diagnosis of breast cancer can be used to evaluate the invasion of tumors and provide important information for accurate diagnosis and treatment. Statistics the number of mitosis cells in breast cancer is one of the important indicators of breast cancer division. In this paper, we summarized the current mainstream methods of mitosis cells detection. These methods are mainly implemented based on deep learning, and discussing the results of some of the methods, comparison and evaluation. At last, through the review of the research methods in this field, the existing breast cancer research methods have been summarized, and the future developments are prospected.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"213 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134439043","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-16DOI: 10.1109/ICCWAMTIP56608.2022.10016537
Li Wei
With the advent of the Big Data era, anomaly detection has become an important tool for screening the validity of data. Many well-established distance-based or correlation-based anomaly detection methods are widely used for various structured and feature-based datasets with the increasing size of data. However, different method strategies have different focuses, leading to large deviations in anomaly detection results for the same dataset using different methods, which poses a great challenge to anomaly detection research. In this paper, a new strategy is proposed for anomaly detection using integrated methods. By using a two-stage process of the sliding window aggregation method, the strategy uses a multi-model anomaly scoring method and a uniform quantitative criterion filtering to obtain a suitable anomaly scoring.
{"title":"A Chaos-Based and Ensembled Method for Outlier Detection","authors":"Li Wei","doi":"10.1109/ICCWAMTIP56608.2022.10016537","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016537","url":null,"abstract":"With the advent of the Big Data era, anomaly detection has become an important tool for screening the validity of data. Many well-established distance-based or correlation-based anomaly detection methods are widely used for various structured and feature-based datasets with the increasing size of data. However, different method strategies have different focuses, leading to large deviations in anomaly detection results for the same dataset using different methods, which poses a great challenge to anomaly detection research. In this paper, a new strategy is proposed for anomaly detection using integrated methods. By using a two-stage process of the sliding window aggregation method, the strategy uses a multi-model anomaly scoring method and a uniform quantitative criterion filtering to obtain a suitable anomaly scoring.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133973791","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-16DOI: 10.1109/ICCWAMTIP56608.2022.10016530
Hongjin Ren, Jinbiao Kang, Kecheng Zhang
Building smart cities in the new era depend heavily on traffic flow analysis, forecast, and management. How to integrate time series and spatial data is a crucial difficulty for anticipating traffic patterns in a smart city. An evident flaw in the existing GCN-based approach is that it is unable to collect non-adjacent but related spatial information because the adjacency matrix only contains the original topological spatial information. In this work, we create a brand-new kind of adjacency matrix that includes both prospective spatial relationships and unique spatial properties using a cutting-edge data-driven methodology. Furthermore, we develop a high-accuracy Spatio-Temporal Graph-TCN Neural Network, called ST-GTNN, for traffic flow prediction. The graph spatial attention layer and the channel attention layer are specifically used to be aware of spatial features, whereas the TCN layer and the temporal attention mechanism are used to fit temporal interactions. Experiment results on two real datasets show that our proposed ST-GTNN outperforms existing methods in terms of prediction accuracy.
{"title":"Spatio-Temporal Graph-TCN Neural Network for Traffic Flow Prediction","authors":"Hongjin Ren, Jinbiao Kang, Kecheng Zhang","doi":"10.1109/ICCWAMTIP56608.2022.10016530","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016530","url":null,"abstract":"Building smart cities in the new era depend heavily on traffic flow analysis, forecast, and management. How to integrate time series and spatial data is a crucial difficulty for anticipating traffic patterns in a smart city. An evident flaw in the existing GCN-based approach is that it is unable to collect non-adjacent but related spatial information because the adjacency matrix only contains the original topological spatial information. In this work, we create a brand-new kind of adjacency matrix that includes both prospective spatial relationships and unique spatial properties using a cutting-edge data-driven methodology. Furthermore, we develop a high-accuracy Spatio-Temporal Graph-TCN Neural Network, called ST-GTNN, for traffic flow prediction. The graph spatial attention layer and the channel attention layer are specifically used to be aware of spatial features, whereas the TCN layer and the temporal attention mechanism are used to fit temporal interactions. Experiment results on two real datasets show that our proposed ST-GTNN outperforms existing methods in terms of prediction accuracy.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123162238","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-16DOI: 10.1109/ICCWAMTIP56608.2022.10016589
Zhao Zekuan, He Chunlin
In the daily life of the city, the normal operation of underground drainage pipes is a necessary condition to ensure the normal life of residents. However, with the increase of the service life of the drainpipe and the improvement of the function of water transmission and sewage, it is particularly important to evaluate the state of the drainpipe. However, the traditional pipe network detection methods such as CCTV and periscope detection are not only inefficient but also cost high. Nowadays, the Object Detection technology is becoming more and more mature, and the application of image detection technology to the defect detection of drainage pipe network is also a hot research direction. Therefore, an improved YOLOv5 Object Detection method was selected in this paper to realize the defect detection of drainage pipe network. In addition, in order to better complete the detection task in the complex image background of the waterway, the multi-head attention mechanism was incorporated into the backbone network of YOLOv5, and the FPN+PAN structure of YOLOv5 was replaced by BiFPN structure. Finally, through simulation experiments, the Precision(P) of the YOLOv5-TB model used in this paper reached 93.1%, the Recall(R) reached 85.5%, and the Mean Average Precision(mAP) reached 88.4%. Moreover, the mAP increased by 1.1% on the basis of YOLOv5. The simulation results show that the model used in this paper can well complete the task of drainage network defect detection.
在城市的日常生活中，地下排水管道的正常运行是保证居民正常生活的必要条件。然而，随着排水管使用寿命的增加和输水、排污功能的提高，对排水管的状态进行评估就显得尤为重要。然而，传统的管网检测方法，如闭路电视和潜望镜检测，不仅效率低，而且成本高。如今，目标检测技术日趋成熟，将图像检测技术应用于排水管网缺陷检测也是一个热门的研究方向。因此，本文选择一种改进的YOLOv5 Object Detection方法来实现排水管网的缺陷检测。此外，为了更好地完成航道复杂图像背景下的检测任务，在YOLOv5的骨干网中加入了多头注意机制，并将YOLOv5的FPN+PAN结构替换为BiFPN结构。最后，通过仿真实验，本文所采用的YOLOv5-TB模型的Precision(P)达到93.1%，Recall(R)达到85.5%，Mean Average Precision(mAP)达到88.4%。此外，mAP在YOLOv5的基础上增加了1.1%。仿真结果表明，本文所采用的模型能够很好地完成排水管网缺陷检测的任务。
{"title":"Research on Defect Detection Method of Drainage Pipe Network Based on Deep Learning","authors":"Zhao Zekuan, He Chunlin","doi":"10.1109/ICCWAMTIP56608.2022.10016589","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016589","url":null,"abstract":"In the daily life of the city, the normal operation of underground drainage pipes is a necessary condition to ensure the normal life of residents. However, with the increase of the service life of the drainpipe and the improvement of the function of water transmission and sewage, it is particularly important to evaluate the state of the drainpipe. However, the traditional pipe network detection methods such as CCTV and periscope detection are not only inefficient but also cost high. Nowadays, the Object Detection technology is becoming more and more mature, and the application of image detection technology to the defect detection of drainage pipe network is also a hot research direction. Therefore, an improved YOLOv5 Object Detection method was selected in this paper to realize the defect detection of drainage pipe network. In addition, in order to better complete the detection task in the complex image background of the waterway, the multi-head attention mechanism was incorporated into the backbone network of YOLOv5, and the FPN+PAN structure of YOLOv5 was replaced by BiFPN structure. Finally, through simulation experiments, the Precision(P) of the YOLOv5-TB model used in this paper reached 93.1%, the Recall(R) reached 85.5%, and the Mean Average Precision(mAP) reached 88.4%. Moreover, the mAP increased by 1.1% on the basis of YOLOv5. The simulation results show that the model used in this paper can well complete the task of drainage network defect detection.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125325204","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-16DOI: 10.1109/ICCWAMTIP56608.2022.10016478
M. Ayenew, Hang Lei, Xiaoyu Li, Kulla Kekeba, Maregu Assefa, Abebe Tegene, S. Muhammed, H. Leka
Renewable electric energy with reliable supply contributes to society, the economy, and the environment. Careful management of electric power from the consumers’ side is crucial on top of stable production, transmission, and distribution systems for reliable consumption. Electric power supply and consumption have been problematic in urban areas of Ethiopia, where frequent power interruptions come from overloaded transmission and distribution systems. In this paper, we proposed a focused Demand Side Management approach for improving reliable consumption in Addis Ababa. We used data analytics and machine learning (K-mean and long and short-term memory) approaches to understand the data, identify potential customers, and predict the aggregate substation load. We identified intermediate and supper-peak demand hours and potential customers for price-based demand load shifting management. Further, the analysis shows that an increase in electric prices at peak hours causes a reduction in electric demand. Consequently, it reduces distribution load and improves reliability.
{"title":"Data Analytics and Machine Learning for Reliable Energy Management: A Case Study","authors":"M. Ayenew, Hang Lei, Xiaoyu Li, Kulla Kekeba, Maregu Assefa, Abebe Tegene, S. Muhammed, H. Leka","doi":"10.1109/ICCWAMTIP56608.2022.10016478","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016478","url":null,"abstract":"Renewable electric energy with reliable supply contributes to society, the economy, and the environment. Careful management of electric power from the consumers’ side is crucial on top of stable production, transmission, and distribution systems for reliable consumption. Electric power supply and consumption have been problematic in urban areas of Ethiopia, where frequent power interruptions come from overloaded transmission and distribution systems. In this paper, we proposed a focused Demand Side Management approach for improving reliable consumption in Addis Ababa. We used data analytics and machine learning (K-mean and long and short-term memory) approaches to understand the data, identify potential customers, and predict the aggregate substation load. We identified intermediate and supper-peak demand hours and potential customers for price-based demand load shifting management. Further, the analysis shows that an increase in electric prices at peak hours causes a reduction in electric demand. Consequently, it reduces distribution load and improves reliability.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117247716","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：