Unsupervised Bump Hunting Using Principal Components
This is an incremental improvement for statistical data analysis in mode hunting.
The paper tackled the problem of improving mode hunting in multivariate data by investigating when principal component rotation of predictors enhances mode estimators, developing a fast version of PRIM for theoretical analysis and demonstrating improved estimators through simulations.
Principal Components Analysis is a widely used technique for dimension reduction and characterization of variability in multivariate populations. Our interest lies in studying when and why the rotation to principal components can be used effectively within a response-predictor set relationship in the context of mode hunting. Specifically focusing on the Patient Rule Induction Method (PRIM), we first develop a fast version of this algorithm (fastPRIM) under normality which facilitates the theoretical studies to follow. Using basic geometrical arguments, we then demonstrate how the PC rotation of the predictor space alone can in fact generate improved mode estimators. Simulation results are used to illustrate our findings.