Transactions Papers Feedforward Neural Structures in Binary Hypothesis Testing
Stella N. Batalama, A.G. Koyiantis, Panayota Papantoni-Kazakos, Demetrios Kazakos · 1993
In this paper, two feedforward neural structures are considered, whose objective is binary hypothesis testing. The first structure, named FFS1, is a tandem structure, while the second structure, named FFS2, involves cumulative feedforward feedback. Both parametric and robust designs for the two struc- tures are considered and analyzed in terms of induced false alarm and power probabilities. The inferiority of the FFSl is rigorously proven in terms of the rate with which the induced power probability increases with respect to the number of the neural elements. Asymptotic results are presented, as well as numerical results, with emphasis on the Gaussian and location parameternominal hypotheses model. Learning algorithms for the parameters involved in the robust network designs are discussed as well.