With respect to anomaly detection and setting inferred fields for a float type scalar encoded input, what is the difference in the quality of anomaly scores on using SDRClassifier versus CLAClassifier?
What kind of situation/use case calls for the usage of one over the other?