Matching algorithms or classifiers determine if a previously enrolled instance matches an observed instance based on some rules. They return a decision, which consists of three possible answers: match, non-match, and unclassified. A classifier assigns a class label to a sample and then checks the new instance with a sample one. Or, the classifier is trained with example instances so that it learns what class label should be applied to future unknown instances. Classifiers are based on statistical, probabilistic, and decision rules. In applying classifiers, the most important issue is finding the matching rates. Two important rates are the false acceptance rate (FAR) and the false rejection rate (FRR). In this work, we determine the FAR and FRR for the Hotelling''s two-sample T2 algorithm applied to the application of matching electronic fingerprints of radio frequency identification (RFID) tags in the presence of simulated noise. The algorithm is found to be a robust classifier for this application.