ANALYZING ECONOMIC STRUCTURE AND COMPARING THE RESULTS OF THE PREDICTED ECONOMIC GROWTH BASED ON SOLOW, FUZZY-LOGIC AND NEURAL-FUZZY MODELS
研究了伊朗经济增长的供给侧因素，比较了索洛、模糊逻辑和神经模糊模型的预测效果，发现神经模糊模型预测更优。
Investigating the factors effective on economic growth is of great importance for most economists. Although lots of studies have been done on economic growth in the world, it has less been regarded in Iran. In this article, by estimating growth regression, we attempt to investigate the supply side of economic growth in Iran. Then we compare the predictive results of Fuzzy-logic, Neural-Fuzzy and Solow models. The results show that there was negative significant relationship (i.e.–0.035) between unstable policy and economic growth rate in Iran during investigation period (1959–2001). In this model, the effect of expenses used by government is positive (i.e. 0.01). Furthermore, the estimated results of long term relationship show that the variable coefficients of capital, labor power, exportation, and inflation are 0.319, 0.016, 0.001, and–0.001, respectively. And also by comparing the predictive results of models for the average percent of annual growth, it is predicted that the average percent of Solow, Neural-Fuzzy, and Fuzzy-logic models are 7.17%, 5.92%, and 6.46% for 2002–2006, respectively. Evaluation of results from the models on the basis of criteria shows that model Neural-Fuzzy predicts better than Fuzzy-logic and Solow models. In other words, forecasting by the model Neural-Fuzzy is recommended.