IMPROVEMENT TO A NEURAL-NETWORK CLOUD CLASSIFIER

Examination of various feature selection algorithms has led to an impr ovement in the performance of a probabilistic neural network (PNN) clo ud classifier. These algorithms reduce the number of network inputs by eliminating redundant and/or irrelevant features (spectral, textural, and physical measurements). One such algorithm, selecting 11 of the 2 04 total features, provides a 7% increase in PNN overall accuracy comp ared to an earlier version using 15 features. This algorithm employs t he same search procedure as before, but a different evaluation functio n than used previously, which provides a similar bias to that of the P NN classifier. Noticeable accuracy improvements were also evident in i ndividual cloud-type classes.