Título: Preliminary Testing and Analysis of an Adaptive Neural Network Training Kalman Filtering Algorithm
Autores: Silva, Jaime Augusto da; Rios Neto, Atair
Resumo: An adaptive neural network training Kalman filtering algorithm is implemented. The XOR problem and a benchmark problem for diagnosis of breast cancer are used for testing and analysis of the algorithm behavior. Results show that the algorithm performs well on both problems having the desirable characteristics of beeing simple to implement, with parallel processing features and good numerical behavior due to the adaptive distribution of learning.
Artigo em pdf: 4cbrn_057.pdf
Arquivo BibTex: 4cbrn_057.bib