id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_ujtg6wuf65dkbcqfvnwz2jnys4	Ivan Maric	Optimization of self-organizing polynomial neural networks	2013	21	.pdf	application/pdf	8695	1431	74	trained by the group method of data handling (GMDH) algorithm is a partial optimization of model weights as procedures for the correction of the flow rate error and for the estimation of optimal set of input parameters (Table Optimal Set of input Parameters for Natural Gas Flow Rate Correction Factor Modelling corresponding SOPNN model (Fig. 5) is limited to maximum 27 polynomial nodes with the maximum ET, Average RRSE of the correction factor for 10 test data sets when approximated by MLP and SOPNN models The SOPNN represents the 27-node self-organized model (Fig. 5) trained by the GMDH algorithm & Hauser, 2010) on the test data set, obtained by the non-optimized GMDH model (SOPNN) and by the same sinusoids from y5 MSO by using the same model (Fig. 8) generated and optimized on y5 training data. SOPNN modeled and optimized by using (y5, n=1,...,400) 50-dimensional training data set.	./cache/work_ujtg6wuf65dkbcqfvnwz2jnys4.pdf	./txt/work_ujtg6wuf65dkbcqfvnwz2jnys4.txt