id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_krmvgatlr5ce3mvlxofc3t3e2i	I. HATZILYGEROUDIS	CONSTRUCTING MODULAR HYBRID RULE BASES FOR EXPERT SYSTEMS	2001	19	.pdf	application/pdf	7863	894	73	difficulty of the adaline unit to classify non-separable training examples, the notion of of 'close' examples are produced from the initial training set and a copy of the neurule for each That is, in case of failure, we produce two subsets of the initial training set of the involved neurule, which contain 'close' success examples and train a copy of the s neurules produced from empirical (training) data (see Section The algorithm for constructing a hybrid rule base from empirical (training) data is each initial neurule using its training set and produce the corresponding produce as many initial neurules as the different intermediate and output variablevalue pairs specified. that the output variable can take three possible values, we need three initial neurules, The training sets of the initial neurules are extracted from the empirical data Initial training sets for the contact lenses example The neurule base for the contact lenses example 	./cache/work_krmvgatlr5ce3mvlxofc3t3e2i.pdf	./txt/work_krmvgatlr5ce3mvlxofc3t3e2i.txt