Discriminative Fuzzy C-Means as a Large Margin Unsupervised Metric Learning Algorithm

Moslehi, Zahra; Taheri, Mahsa; Mirzaei, Abdolreza; Safayani, Mehran

doi:10.1109/tfuzz.2018.2836338

Cited by 9 publications

(1 citation statement)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the traditional K-means [17] and fuzzy C-means [18] clustering are more sensitive to the setting of the initial value, easy to converge to the local extreme point, and the number of clusters needs to be given in advance. With the continuous development of artificial intelligence, swarm intelligence algorithms are gradually being used to solve various clustering problems [19][20][21][22][23][24][25].…”

Section: Introductionmentioning

confidence: 99%

Underdetermined blind source separation of radar signals based on genetic annealing algorithm

Jin¹,

Luo²,

Li³

et al. 2021

The Journal of Engineering

View full text Add to dashboard Cite

Aiming at the problem that the traditional underdetermined blind source separation (UBSS) algorithm does not have high accuracy in estimating the mixing matrix and the number of source signals cannot be determined, a genetic annealing UBSS method with source number estimation function is proposed. This method first performs a short-time Fourier transform on the observed signal, and then performs single-source point detection and normalization processing to transform linear clusters into dense clusters. In view of the strong global search ability of the genetic algorithm (GA) and the good local optimization ability of the simulated annealing algorithm, the GA and simulated annealing algorithm are combined, and the number of source signals is determined according to the BWP (Between-Within Proportion, BWP) index. After estimating the mixing matrix, a modified subspace projection method is used to reconstruct the source signal. The simulation results of four kinds of radar signals show that in the case of dual receiving channels and three receiving channels, compared with the traditional method, the similarity coefficients of the proposed algorithm are greater, which can effectively improve the mixing matrix estimation accuracy, and can accurately determine the number of source signals.

show abstract