Analyzing and Identifying the Effect of Fluctuations of Oil, Gold and Currency Variables on the Trade Balance in Iran Using the Combined Approach of Generalized TOPSIS and Fuzzy Regression

In this study, analysis and identification of oil, gold and currency relations on the trade balance in Iran using the new hybrid approach of fuzzy regression and multi-criteria decision making. For this purpose, five fuzzy regression models FERRARO, M_ESTIMATION, LA_FUZZY, WEIGHTED FUZZY RWGRWSSION and FUZZY_OLS used for this purpose three goodness of fit criteria, in the form of options and indicators of a multi-criteria decision problem as input, were used for the annual data during the period of 1357-1999. TOPSIS decision making models were organized and then the above models were ranked and compared. To run the model, Python 3.9 software is used, and among the above models, the fuzzy weighted model has been introduced as the selected model. According to the results, among the variables of oil, gold and foreign exchange, the exchange rate variable has the most impact on the country's trade balance during the period under review. other results showed that the use of regression approaches with various logics in the design of the error function produces more accurate estimates of the coefficients. The range of minimum and maximum effect of exchange rate on trade balance was extracted as (1.474 and -1.382). These fuzzy cumulative regression coefficients, which are a combination of (regression coefficients, correlation coefficients, entropy, absolute regression coefficients) of variables, show that the priority of influencing the trade balance category is: exchange rate, oil price and then gold price.