Description

A data transformation based on the characteristics of semiconductor chips on a wafer (left) reveals the hidden correlation among the chips, which, when visualized in the transformed space (right), demonstrates a colorful heart.

Kernel PCA (principal component analysis) based on the test measurements of the semiconductor chips on a wafer maps the samples into a transformed space, in a unique way. The samples were colored to visualize the mapping. The analysis is based on industrial production test data and could help identify abnormalities in the chip population.