Improving out-of-sample forecasting is one of the main issues in financial research. Previous studies have achieved this objective by increasing the number of input variables or changing the kind of input variables. Changing the forecasting model is another possible approach to improve out-of-sample forecasting. Most researches have focused on linear models, while few have studied nonlinear models. In the present study, we have reduced the number of variables and at the same time applied a nonlinear forecasting model. Oil prices have been used as predictors to predict return by application of a new artificial neural network nonlinear model named Deep Learning and its comparison with OLS and ANN methods. Results indicate that the applied non-linear model has higher accuracy compared to historical average model, OLS and ANN. It also indicates that out-of-sample prediction improvement does not always depend on high input variables numbers. On the other hand when using a smaller number of input variables, it is possible to improve this forecasting capability by changing the model and applying nonlinear models.

The present study aims to identify the variables affecting the price of cement and provide in- and out-of-sample forecasts using monthly data from 2019:03 to 2023:02 and a vector autoregression (VAR) model. In this regard, according to the results of the Johansen-Juselius cointegration test, the long-term relationship between the variables was confirmed. Then, the long-term and short-term models were estimated, and the error correction coefficient was -0.0858. In the next stage, based on the results of impulse response functions, exchange rate and construction input price index shocks (respectively with a positive effect of 8 and 6.8 percent in the standard form) have been more effective than other model variables on the fluctuations of cement price. The results of variance decomposition also showed that the construction input price index, energy cost, and exchange rate are important in explaining the price of cement. Finally, we estimated in-sample and out-of-sample forecasts. Based on the forecast evaluation criteria, our founding research model can accurately predict the price trend of cement.

The abundant and numerous uses of steel in various industries have turned it into a strategic commodity, and its price has always been a concern of industrial owners. Therefore, access to accurate forecasts of the price trend of steel and its products is important. Therefore, in the present study, while identifying the determinant variables for the price of steel, an out-of-sample forecast for 2023:10 to 2024:06 has been made using the vector autoregression (VAR) model. At first, the results of the Johansen-Juselius cointegration test confirmed the long-term relationship. Also, the error correction coefficient (ECM) was -0.0842. We analyze impulse response functions. The unofficial exchange rate, and industrial producer price index (base metal manufacturing sub-group) (respectively with a positive effect of 6.8% and 6.5% percent in the standard form) have been more effective than other model variables on the fluctuations of steel price. In addition, the results of variance decomposition showed that the industrial producer price index (16.18%), and unofficial exchange rate (11.7%) after the price of steel itself had more effect on the price fluctuations of this product than other variables. Finally, we estimate the out-of-sample forecast. The price of steel is forecasted from 302,445 Rials in October 2023 to 321,552 Rials in June 2024 (with a 5% increase). Based on the forecast evaluation criteria, our model can accurately forecast the price trend of steel. According to the results, the most important policy recommendations include updating the technologies, equipment, and machinery used in the production of steel products to develop the production of steel products, saving energy consumption and production costs, adopting appropriate monetary and foreign exchange policies, And increasing exports by focusing on marketing and sales of steel products (to prevent raw sales), through the development of production, increasing the variety, quality, and durability of manufactured steel products can be provided

Considering the importance of rice and the significant share of its consumption in the household basket, the correct policy to control its price in the market will play a chief role in the well-being and food security of households. Therefore, checking the price trend and providing price forecasts for this product is significant. Hence, the present study aims to identify the influencing variables and investigate their impact on the price of rice and provide out-of-sample forecast (2023:05-2024:03) using vector autoregression method. Estimation of two models for Iranian Caspian rice and foreign Thai rice; using up-to-date and monthly data; Also, price forecasting of these two rice is one of the innovations of this study. Johanson-Jusilius's cointegration test confirmed the existence of a long-term relationship between variables. Then, the long-term and short-term (VECM) models were estimated and the error correction coefficients for the first and second models were -0.300 and -0.309, respectively, which are significant at the 1% level. Next, according to the results of impulse response functions, the shocks exchange rate and the price of the substitute product in model 1, and the shocks price of the substitute product and the global price index in model 2, have been more effective than other variables on the price fluctuations of Caspian and Thai rice. Finally, we estimated out-of-sample forecasts (with two currency scenarios). Based on the prediction evaluation criteria, the research models can accurately predict the price trend of Iranian and foreign rice.

Forecasting of crude oil price plays a crucial role in optimization of production، marketing and market strategies. Furthermore، it plays a significant role in government’s policies، because the government sets and implements its policies not only according to the current situation but also according to short run and long run predictions of important economic variables like oil price. The main purpose of this study is modeling and forecasting spot oil price of Iran by using GARCH model and A Gravitational Search Algorithm. Performed forecasts of this study are based in static and out-of-sample forecasting and each subseries data is divided in to two parts: data for estimation and data for forecasting. The forecast horizon is next leading period and its length is one month. In this study the selected models for forecasting spot oil of Iran are GARCH (2،1) and a Cobb Douglas function which is functional of prices of 5 days ago. Finally، the performances of these models are compared. For comparison of these models MSE، RMSE، MAE، and MAPE criteria are used and the results indicate that except in MAPE criterion، the mentioned criteria are smaller for GARCH model in comparison to GSA algorithm.

This paper discusses the problems of short-term forecasting of cryptocurrency time series using a supervised machine learning (ML) approach. For this goal, we applied two of the most powerful ensemble methods including Random Forests (RF) and Stochastic Gradient Boosting Machine (SGBM). As the dataset was collected from daily close prices of three of the most capitalized coins: Bitcoin (BTC), Ethereum (ETH) and Ripple (XRP), and as features we used  past price information and technical indicators (moving average). To check the effectiveness of these models we made an out-of-sample forecast for selected time series by using the one step ahead technique. The accuracy rate of the forecasted prices by using RF and GBM were calculated. The results verify the applicability of the ML ensembles approach for the forecasting of cryptocurrency prices. The out of sample accuracy of short-term prediction daily close prices obtained by the SGBM and RF in terms of Mean Absolut Percentage Error (MAPE) for the three most capitalized cryptocurrencies (BTC, ETH, and XRP) were within 0.92-2.61 %.

Appropriate credit scoring assessment assists financial institutions on loan pricing, determining amount of loan, loan risk management, reduction of default risk and increase in debt repayment. The purpose of this study is to estimate a credit scoring model for the agricultural loans in Kohgiloye & Bovair Ahmad Province. The logistic regression with the two estimation methods (classic and Bayesian) are used to construct the credit scoring models as well as to predict the borrower’s creditworthiness and default risk. Furthermore, Bayesian method was compared with the classical estimation methods. The Limdep and MLwiN softwares are used to estimate models by Classic and Bayesian approaches, respectively. Data were collected from 110 farmers in Kohgiloye & Bovair Ahmad Province in 2007. Results of the Bayesian method indicated that variables such as education, value of assets and age of farmers have positive effects, whereas borrowing from others; loan type, total debet to assets ratio and the duration of bank-borrower relationship have negative effect as important factors in determining the creditworthiness of the borrowers. The results also show that a higher value of assets implies a higher creditworthiness and a higher probability of a good loan. However, the negative sign found on the duration of bankborrower relationship, which contradict with the hypothesized sign, suggest that the borrower who has a longer relationship with the bank has a higher probability to default on debt repaymen. The overall prediction accuracy of the Bayesian credit scoring models is 90.91% and 89.91% in-sample and out-of-sample forecast, respectively, and is higher than the classic model on out-of-sample forecast. Thus, when the expected loss of misclassification are computed and compared, the results indicate that the misclassification cost of the Bayesian method is the best credit scoring model with the lowest misclassification costs. In summary, the empirical results in this study support the use of the Bayesian method in classifying and screening agricultural loan applications.

In this paper, Using CPI data from 1990 to 2011, it is showed that Iranian inflation series has been encountered with structural breaks. Then, it is showed that time-varying parameter models can explain the behavior of Iranian inflation, but nonlinear model cannot. Also, investigating the performance of out-of-sample forecasting shows that the performance of time-varying parameter model is slightly better than benchmark AR model in all forecast horizons, although this difference is not significant, but the nonlinear model in all forecast horizons does not have better performance than our benchmark model. So, although modeling inflation by time-varying models can explain the behavior of inflation, but it cannot help forecast inflation.

This study provides a dynamic estimation of Capital Asset Pricing Model in optimal dynamic household decisions during its life cycle based on Dynamic Conditional Correlation (DCC) and State-Space Models. In this regard, for daily time series data of Vehicle and parts manufacturing stock price for the period of 2006 to the end of sixth month of 2011, dynamic estimation of Beta-index time series which obtained from DCC and State-Space models has been compared with static linear regression model. The results show that the DCC model has minimum error in out-of-sample forecasting of stock returns based on Root Mean Square Error (RMSE) criterion. However, state-space model has minimum error for in-sample forecasting based on RMSE criterion.

Globalization, the process of considerable increase in international trade, global exchanges and markets’ integration as a fundamental characteristic, are emerging inevitably. Investigating the way in which this process affects economic variables, can be a guidance of decision making for policy makers. Considering structural economic differences between urban and rural societies in Iran and using 1350-1386 Iran’s economy dataset, first we compared the efficiency of VAR and VEC models with artificial neural network (ANN) approach in forecasting measure of income distribution inequality of urban societies and finally the best model (ANN) has used as an out-of-sample forecasting tool in different designed scenarios from 1387 to 1395. Choosing ANN model, decrease in urban societies’ income inequality during globalization process, is the main result.