The Prediction of E-Money Circulation: Backpropagation with Genetic Algorithm Adoption

Budianita, E.; Okfalisa, O.; Assiddiki, M. R.

Digital transformation forces the utilization of e-money during the economic transaction. Behind its advantages, e-money has been influenced by the inflation rate, thus accelerating the country's money circulation. Moreover, the fragile Covid-19 economy triggers each country's need to anticipate the circulation of e-money to deter future inflation. Therefore, this paper deployed the Backpropagation approach integrated with the Genetic Algorithm to forecast the dissemination of e-money in Indonesia by exploiting time-series Bank Indonesia (BI) data from January 2009 to December 2019. Here, 120 data with 12 variables are considered to thoroughly predict the Year 2020 circulation focusing on the previous 12 months. This study reveals that e-money circulation in Indonesia is increasing monthly in 2020. The testing result shows that the lowest mean square error (MSE) is found at 0.000035 for data training division at 90%:10%, learning rate parameter at 0.8, the combination of crossover probability and mutation at 0.4:0.6, and the total generation and population at 350 and 200, respectively. In a nutshell, Backpropagation with a Genetic Algorithm has been expected to a successful outcome for e-money circulation and provides large values compared with actual data and original BPNN. © 2021 IEEE.