H. Steigstra

TL;DR: An algorithm called SOLOMON is presented for classification of patterns in multi-dimensional space by constructing a statistical model based on multivariate analysis of the classes under study by using multi-inductive component analysis.

TL;DR: A powerful transformation technique in pattern recognition is presented as an alternative to principal component analysis and its properties include automatic feature selection, presentation of physically interpretable results, and elucidation of hidden relationships between variables.

TL;DR: The software package comprising the MICA transformation technique and the SOLOMON classification program is extended to include an algorithm for estimating and supplementing missing data, which is compared with that of known classification techniques on simulated data sets.

TL;DR: A new technique for graphical representation is given as an alternative for PCA, with main advantages are the inherent feature selection, a better understanding of the result of the transformation and its simplicity.