Artificial Neural Networks with Radial-Basis Transmission Functions for Modelling Economic Phenomena and Processes
Purpose: The paper presents the possibility of using artificial neural networks (ANN) with radial-basis transmission function (RBF) for modeling economic phenomena and processes. Design/Methodology/Approach: The basic characteristics and parameters of an ANN with RBF are shown and the advantages of using this type of ANN for modeling economic phenomena and processes are emphasized. Using an ANN with RBF, together with official statistics for 2010-2017, the modeling of the influence caused by work efficiency indicators of the customs authorities of Ukraine on the indicators of economic security of Ukraine was carried out. These eighteen indicators of economic security of Ukraine, which comprehensively characterize the economic status of the country in terms of production, social, financial, food, transport, energy, and foreign economic security, were chosen as the most informative indicators. Findings: The results of the study showed that Artificial neural networks with Radial-basis transmission function well describe the trend of changing state economic security indicators under the influence of changing performance indicators of customs authorities. This allows us to recommend this type of artificial neural networks for analysis, evaluation and forecasting of economic phenomena and processes. Practical Implications: The influence of the efficiency of fiscal bodies on the main indicators of economic security of the state is proved. The analytical and prognostic properties of ANN with RBF for estimating the relationships between economic phenomena and processes, under the conditions of non-stationary processes and nonlinearity of relationships between indicators, are shown. Originality/Value: The results obtained showed good analytical and prognostic properties of an ANN with RBF when modeling the impact of customs authorities' performance on the state's economic security indicators.