Smart selection of indirect parameters for DC-based alternate RF IC testing

Brownian distance correlation-directed search: A fast feature selection technique for alternate test

“…In addition, we will also comment the use of Principal Component Analysis (PCA) which is not strictly speaking a feature selection method but rather a dimensionality reduction technique. Despite of this, the reason why we include it into this study is because it is a widely-used preprocessing step in machine-learning and has actually been applied to Alternate Test [18].…”

Efficient Production Binning Using Octree Tessellation in the Alternate Measurements Space

“…It is quite clear that a set of alternate measurements to be used for binning purposes has to satisfy two main properties: (1) the measurements need to reflect circuit's performances variability in order to allow the binning to be performed efficiently and (2) an adequate set of measurements should not be redundant to avoid incurring in extra binning costs [31]- [33]. The first condition is achieved by means of a sensitivity analysis of candidate alternate measurements.…”

Special session: Hot topics: Statistical test methods

“…Many factors influence this efficiency such as the choice of adequate indirect parameters, the choice of the prediction model, the order of mapping between indirect measurements and device specifications, or the size and composition of the training set. The objective of this talk is to discuss various statistical methods for the choice of adequate indirect measurements [4]- [7], pertaining to both filter and wrapper categories in the field of feature selection. Their impact on alternate test efficiency will be evaluated in terms of model and prediction accuracy by using classical metrics such as average and maximal errors, but also in term of prediction reliability by introducing a new metric called Failing Prediction Rate (FPR).…”