id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_hwz54ezyk5eq3arb5vhhhhhily	Antonio Maratea	Record linkage of banks and municipalities through multiple criteria and neural networks	2020	20	.pdf	application/pdf	6570	710	58	multiple similarity criteria and a Neural Network classifier is proposed: starting from Keywords Record Linkage, Entity resolution, Neural networks, Feature extraction, Deduplication 4. Classification: the similarity vector obtained from each pair of records within the same a set of ensemble learning methods combining multiple base classifiers, including a Figure 2 Example of the similarity vector obtained comparing two records using four similarity functions (please see Table 2 for attribute mapping). The training data-set, in the format (feature,label) is generated based on the candidate • feature: is the similarity vector obtained comparing the records of the pair; Since non-matching record pairs are more than matching ones, the training data Figure 8 Results for multiple criteria similarity with threshold θ = 0.63, weighted. Figure 9 Results of the Levenshtein test with MLP classifier. then a classifier based on multiple criteria and Neural Networks has been proposed in Data matching: concepts and techniques for record linkage, entity	./cache/work_hwz54ezyk5eq3arb5vhhhhhily.pdf	./txt/work_hwz54ezyk5eq3arb5vhhhhhily.txt