Technique of cluster analysis application for electronic components classification and equipment reliability improvement

The paper deals with the technique of cluster analysis application for experimental facilities classification and quality and reliability improvement of electronic equipment. K-means clustering method is used for carrying-out of cluster analysis. Studies were conducted on a sample of chips 765LN2. The centering of initial data (informative parameters) was used. By means of hierarchical classification the vertical dendrogram of cluster analysis was obtained. The dispersive analysis of results was carried out. It allowed to determine the importance of distinctions between clusters. It is stated that the cluster analysis is effective in case when classes of copies are well divided.