Abstract
AbstractSoft independent modeling of class analogy (SIMCA by Wold) is a statistically based class modeling approach, which has numerous modifications. Two basic statistics, the score and orthogonal distances, are utilized in all known SIMCA versions. The difference between the methods is the way they interpret and apply statistics in decision rules. The four most popular SIMCA decision rules, simple (by Massart), alternative (by Wise), combined index (by Joe Qin), and data driven (by Pomerantsev), are described in detail and compared using simulated and real‐world cases. The examples originate from various fields, and what is more important for statistical analysis is that they present different statistical structures. It is shown that depending on the way the decision rule is built, different SIMCA results can be obtained.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
