Soft Computing Methods Applied in Forecasting of Economic Indices Case Study: Forecasting of Greek Unemployment Rate Using an Artificial Neural Network with Fuzzy Inference System

摘要

The unemployment rate is an indicator used by investors to determine the health of the economy. It is also a measurement of the economy growth rate. Greece is a low-productivity economy. Statistics unfortunately cannot be totally taken at face value because earnings of many Greeks and immigrant workers are off-the-books. In addition, the immigrants make up nearly one-fifth of the work force, mainly in unskilled jobs. A 2003-2008 employment action plan included measures for the state to provide some 25,000 part-time jobs, more subsidies and tax incentives, greater social services (especially for women), training programs for the long-term unemployed, rent assistance and breaks unemployment rate to drop. The paper presents an ANFIS forecasting model. The results were presented and compared based on four different kinds of errors: MSE, RMSE, MAE and MAPE. The ANFIS model gives the best results for the case of six gauss membership functions and 250.000 epochs.