MODELO DE PREVISÃO DE DEMANDA PARA ATMS UTILIZANDO REDE NEURAL ARTIFICIAL DO TIPO MULTLAYER PERCEPTRON
Keywords:
Redes Neurais Artificias, Multilayer Perceptron, Reposição de Numerários de ATMs, Previsão de Reabastecimento de ATMsAbstract
The projections of replacement of cash for Automated Teller Machine (ATM) ATM for bank self-service provides to a banking network optimization and greater efficiency in all process of replacement of values in ATMs, as a result providing security, cost reduction, and balance in relation to seasonality. The aim of this work was to develop an Artificial Neural Network (ANN) to estimate the daily withdrawal values, also considering other variables that could influence the seasonality of these movements. The Artificial Neural Network, of the FeedForward Multilayer Perceptron (MLP) type, was trained based on ATM movement data from different location points. In the intermediate and output layers, activation functions of the relu-adam type were used. Thus, the performance of the developed ANN proved to be satisfactory, and it can be considered as a model for the implementation of effective use in the operation of cash replacement in ATMs of the bank networks or in shared ATMs.
Downloads
References
ARORA, N., SAINI, J.R. Approximating Methodology: Managing Cash in Automated Teller Machines using Fuzzy ARTMAP Network. International Journal of Enhanced Research in Science Technology and Engineering. vo.l3, 318-326, 2014.
ATM Marketplace. Global ATM installed base to reach 4M by 2021. https://www.atmmarketplace.com/news/global-atm-installed-base-to-reach-4m-by-2021/. Acesso em: 17 Junho 2020.
BRAGA, A.A. CARVALHO, A.C.P.L.F.; LUDERMINR, T.B. Redes Neurais Artificiais: teoria e aplicações: 2. ed. Rio de Janeiro; Editora LTC, 2007.
COUTINHO, E. et al. Utilização de Técnicas de Inteligência Computacional me Predição de Dados Meteorológicos. Seropédica: Revista Brasileira de Meteorologia, v.31, n.1, 24-36, 2016. https://doi.org/10.1590/0102-778620140115
Deep Learning Book. O Neurônio, Biológico e Matemático. http://www.deeplearningbook.com.br/o-neuronio-biologico-e-matematico. Acessado em: 22 Junho 2020.
Deep Learning Book. Definindo o Tamanho do Mini-Batch. https://deeplearningbook.com.br/definindo-o-tamanho-do-mini-batch/. Acessado em: 19 Julho 2020.
Kaggle Inc. Competitions. https://www.kaggle.com/competitions. Acessado em: 17 Junho 2020.
Kaggle. Data of ATM transactions of XYZ bank. https://www.kaggle.com/nitsbat/data-of-atm-transaction-of-xyz-bank. Acessado em: 10 Abril 2019.
LEONOV, P. et al. The use of artificial intelligence technology in the process of creating an ATM service model: Procedia Computer Science, v.169, 203-208, 2020. https://doi.org/10.1016/j.procs.2020.02.137
MELO, Gedson Santos. Aplicação de Aprendizado de Máquina para Previsão de Fluxo de Caixa em ATMs. 2018. 37f. Monografia para obtenção do grau de Bacharel em Engenharia da Computação - Universidade Federal de Pernambuco, Recife, 2018.
PAIVA, Rodrigo de Carvalho. Modelo de previsão e reposição de numerários em uma rede de caixas eletrônicos. 2006. 68f. Trabalho de Formatura para Conclusão do Curso deEngenharia de Produção - Escola Politécnica da Universidade de São Paulo, São Paulo, 20