In data rich environments, groups of data samples, or signals, can be summarized into trapezoidal fuzzy sets using order statistics. The data stream then appears as a sequence of fuzzy numbers, which are applied to a quadratic discriminant to yield fuzzy numbers that can be ordered to classify the signals. The resultant classifier is a fuzzy quadratic classifier (FQC). If signal-to-noise information is available with each sample, fuzzy order statistics produce a more refined data stream. An example of FQC is given using Slash data where the variation in SNR is assumed to be available to the system. Two fuzzy ranking methods are used to produce a hard and a soft classifier. The soft classifier illustrates how the uncertainty in the data stream captured by modeling the data as fuzzy numbers can be propagated through a discriminant to yield a class membership interval for each signal.