engleski

Forecasting Federal Funds Target Rate By Neural Networks

The aim of the paper was to create a model for predicting US federal funds target rate using neural networks. The model is based on macroeconomic data of USA covering the period from 1959 to 2005. Ten input variables were used, while the output was the federal funds target rate which is used by american federal bank (FED) to ensure the monetary stability in the country. Different neural network architectures were tested using the backpropagation algorithm, and the best neural network model is selected on the basis of test error. The sensitivity analysis is also conducted revealing that the most influencal input variables are the gold price change, and the change of market indices (Dow Jones i S&P500). The modeling results show that the neural network is able to incorporate the relationship among input variables and output. The created model revealed that artificial intelligence methods have great potential in the area of interest rate prediction and could be used for future research in that area.

target interest rate; neural networks; prediction; multi layer perceptron; cross-validation

nije evidentirano