Thông tin chung
  English
  Đề tài NC khoa học
  Bài báo, báo cáo khoa học
  Hướng dẫn Sau đại học
  Sách và giáo trình
  Các học phần và môn giảng dạy
  Giải thưởng khoa học, Phát minh, sáng chế
  Khen thưởng
  Thông tin khác
  Tài liệu tham khảo
  Hiệu chỉnh
 
Số người truy cập: 107,393,334

 Neighboring information exploitation for anomaly detection in intelligent iot
Tác giả hoặc Nhóm tác giả: Thien-Binh Dang, Duc-Tai Le, Moonseong Kim, Hyunseung Choo
Nơi đăng: International Conference on Future Data and Security Engineering; Số: LNISA,volume 13076;Từ->đến trang: 260-271;Năm: 2021
Lĩnh vực: Công nghệ thông tin; Loại: Báo cáo; Thể loại: Quốc tế
TÓM TẮT
The identification of anomalies has become increasingly important for the security of sensory data gathering in the intelligent Internet of Things (iIoT). The current approaches might not be applied to the general cases of anomalies, i.e., both long-term and short-term anomalies, as well as not be suitable with real-time applications such as natural disaster monitoring and early warning systems. To address this challenge, this paper proposes a comprehensive approach, named DWT-PCA Anomaly Detection (DAD) to detect both long- and short-term anomalies. DAD bases on the combination of Discrete Wavelet Transform (DWT) and Principal Component Analysis (PCA) to improve the system performance. In particular, we first utilize the DWT to extract approximation coefficients and detail coefficients from the input data which are capable to highlight long-term and short-term anomalies, respectively. We then exploit the spatial-temporal correlation of neighboring sensors by applying PCA on these coefficients to obtain a high detection accuracy. Numerical experiments based on the real dataset of Intel Berkeley Research reveal that the proposed scheme obtains higher accuracy and a lower false-positive rate on detecting three typical anomalies: drift, noise, and outlier, comparing to existing schemes.
ABSTRACT
The identification of anomalies has become increasingly important for the security of sensory data gathering in the intelligent Internet of Things (iIoT). The current approaches might not be applied to the general cases of anomalies, i.e., both long-term and short-term anomalies, as well as not be suitable with real-time applications such as natural disaster monitoring and early warning systems. To address this challenge, this paper proposes a comprehensive approach, named DWT-PCA Anomaly Detection (DAD) to detect both long- and short-term anomalies. DAD bases on the combination of Discrete Wavelet Transform (DWT) and Principal Component Analysis (PCA) to improve the system performance. In particular, we first utilize the DWT to extract approximation coefficients and detail coefficients from the input data which are capable to highlight long-term and short-term anomalies, respectively. We then exploit the spatial-temporal correlation of neighboring sensors by applying PCA on these coefficients to obtain a high detection accuracy. Numerical experiments based on the real dataset of Intel Berkeley Research reveal that the proposed scheme obtains higher accuracy and a lower false-positive rate on detecting three typical anomalies: drift, noise, and outlier, comparing to existing schemes.
© Đại học Đà Nẵng
 
 
Địa chỉ: 41 Lê Duẩn Thành phố Đà Nẵng
Điện thoại: (84) 0236 3822 041 ; Email: dhdn@ac.udn.vn