A kernel recursive minimum error entropy adaptive filter

Wang, Gang; Yang, Xinyue; Wu, Lei; Fu, Zhenting; Ma, Xiangjie; He, Yuanhang; Peng, Bei

doi:10.1016/j.sigpro.2021.108410

Cited by 22 publications

(5 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This gives an accurate idea of the expected result. Due to the way LMS works, there is no correlation between how ML is learned in theory and how it is learned in practice [67]. These theories aim for convergence, which occurs when repeated learning leads to a single outcome rather than many outcomes.…”

Section: Least-mean-square (Lms)mentioning

confidence: 99%

Machine Learning Solutions for the Security of Wireless Sensor Networks: A Review

Ghadi,

Mazhar,

Shloul

et al. 2024

IEEE Access

View full text Add to dashboard Cite

Energy efficiency and safety are two essential factors that play a significant role in operating a wireless sensor network. However, it is claimed that these two factors are naturally conflicting. The level of electrical consumption required by a security system is directly proportional to its degree of complexity. Wireless sensor networks require additional security measures above the capabilities of conventional network security protocols, such as encryption and key management. The potential application of machine learning techniques to address network security concerns is frequently discussed. These devices will have complete artificial intelligence capabilities, enabling them to understand their environment and respond. During the training phase, machine-learning systems may face challenges due to the large amount of data required and the complex nature of the training procedure. This article focuses on machine learning algorithms used to solve the security issues of wireless sensor networks. This article also focuses on different types of attacks on layers of the wireless sensor network. Moreover, this study addresses several unsolved issues, including adapting machine learning algorithms to accommodate the sensors' functionalities in this network configuration. Furthermore, this article also focuses on open issues in this field that must be solved.

show abstract

Section: Least-mean-square (Lms)mentioning

confidence: 99%

Machine Learning Solutions for the Security of Wireless Sensor Networks: A Review

Ghadi,

Mazhar,

Shloul

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…where C KRMEE and C QKRMEE are the computational complexity of one cycle of the KRMEE and QKRMEE algorithms; C com represents the computational complexity of formulas of the same form in both algorithms; C θ;KRMEE and C θ;QKRMEE are the computational complexity of φ L in (17) [27] and θ L;S in (32). The difference C d;MEE in computational complexity between the KRMEE and QKRMEE algorithms can be expressed as…”

Section: Computational Complexitymentioning

confidence: 99%

“…In the four aforementioned instances, the performance of the QKRGMEE and QKRMEE algorithms is compared with that of the KRLS [8], KRMC [35], KRMEE [27], and KRGMEE [28] algorithms. In Fig.…”

Section: Fig 1 Convergence Curves Under Different Scenariosmentioning

confidence: 99%

“…To further improve the performance of the KAF algorithms, the GMCC is combined with KAF algorithms, as a result of which the generalized kernel maximum correntropy (GKMC) [23] and the kernel recursive generalized maximum correntropy (KRGMCC) [24] are developed. MEE criterion is considered to have better performance than MCC [25], and the kernel MEE (KMEE) algorithm [26] and kernel recursive MEE (KRMEE) algorithm [27] are investigated. Due to its smoothness and stringent positive definiteness, the Gaussian kernel function is generally chosen as the kernel function of the original MEE.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Mixture quantized error entropy for recursive least squares adaptive filtering

Wang

Peng

et al. 2022

Journal of the Franklin Institute

View full text Add to dashboard Cite

“…Entropy [17]- [21] is a more universal adaptive filtering criterion since it quantifies the average information intrinsic in a given PDF. The MSE can be expanded if information is used as an optimality criterion.…”

Section: Introductionmentioning

confidence: 99%

A novel adaptive noise cancellation method based on minimization of error entropy for electrocardiogram denoising

Mousa Saad,

Al-Chaabawi,

H. Hassan

2023

IJEECS

View full text Add to dashboard Cite

<span>In this paper, we use an adaptive method that conforms to the error entropy criterion in order to eliminate noise from cardiac signals electrocardiogram (ECG). In previous works, the mean squared error (MSE) criterion has been used to adaptive noise cancelation of ECG signals, which only has the ability to minimize the second moment of error. The MSE criterion only works optimally on systems with Gaussian noise and stationary signals, so this is not suitable for ECG signals that are non-stationary and have non-Gaussian noise. In contrast, the use of error entropy-based algorithms like minimum error entropy (MEE) is very useful in ECG noise cancelation. The results of the proposed algorithm indicate a significant advantage in terms of signal-to-noise ratio (SNR) and convergence value compared to the algorithm based on MSE criteria.</span>

show abstract

A kernel recursive minimum error entropy adaptive filter

Cited by 22 publications

References 18 publications

Machine Learning Solutions for the Security of Wireless Sensor Networks: A Review

Machine Learning Solutions for the Security of Wireless Sensor Networks: A Review

Mixture quantized error entropy for recursive least squares adaptive filtering

A novel adaptive noise cancellation method based on minimization of error entropy for electrocardiogram denoising

Contact Info

Product

Resources

About