Dimensionality Reduction using SOM based Technique for Face Recognition

Kumar, Dinesh; S., C.; Kumar, S. S.

doi:10.4304/jmm.3.1.1-6

Cited by 4 publications

(1 citation statement)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Firstly, extracts feature vector, calculated node connection matrix, and fault data network model can be established. Network is divided into several data areas by using SOM network [8,9]. According to complex network community structure modularity judges type number of faults for K-means algorithm looking for the K value, then the use of complex network degree of the sample data selection clustering center, finally, finishing clustering analysis by using K-means clustering algorithm.…”

Section: Introductionmentioning

confidence: 99%

Improving K-means Clustering Method in Fault Diagnosis based on SOM Network

Chen¹,

Pan²,

Jiang³

2013

JNW

View full text Add to dashboard Cite

According to the problem of K value and initial cluster centers selection difficult on K-means clustering algorithm, form essential characteristics of the complex network, the fault samples can be abstracted into network nodes, and the connection between samples can be abstracted into edge, and then the network model of fault data can be established .Failure data network model is divided into several regions self-organizing feature map (SOM) network. K value can be determined from the maximum value which is selected in different division result by the use of community modularity at the same time. Complex network node correlation degrees can be calculated to select important nodes as initial clustering center, then by means of K-means clustering realizing clustering diagnosis. This study is applied to rolling bearing clustering the diagnosis examples and has good effect of fault diagnosis

show abstract

Section: Introductionmentioning

confidence: 99%