An adaptive data sorter based on probabilistic neural networks

Abstract

Based oil a d €~~~& e d , pmhh%stic hemal m~w k (PNN)paradigm, a pamlklmtwork can be used to sort data parameters into dasses with hij$-sarting accuracy and low fragmentation. The capabilitiesof the sorter, as applied to W M pulse-data sorting, will be shown. "he PNN implements the statistical Bayesian strategy by computing a joint probability density over all input data parameters te match a group of candidate data classes. The sorting is accomplished by assigning the inputs to the most likely group with highest probability density estimate. Based on test data fiom an ESM system, the PNN has shown significant improvement over conventional rule-based techniques.The parallel computer architecture of PNN is wellsuited for very large-scale integration (VLSX) chip implementation. An 80,000 gate semicustom chip design concept will be described.