forecasting

This study examines the accuracy of Heterogeneous Autoregressive (HAR) models in forecasting the Conditional Value-at-Risk (CVaR) of Exchange-Traded Funds (ETFs) on the Tehran Stock Exchange. The significance of this study stems from the need for better risk management in financial markets, where volatility and jumps significantly affect investment decisions.

Data from nine equity, index, and fixed-income funds were analyzed intraday with high frequency (daily and fifteen-minute intervals) from 2019 to 2022. Three main families of HAR models were evaluated by considering the relevant variables.

The results revealed that models based on second-order variations outperformed others in forecasting Realized Volatility (RV). Additionally, CVaR prediction was more accurate for index funds than for equity and fixed-income funds, with the HARQ model demonstrating superior performance.

Originality/Value: 

This study investigates the application of HAR models in predicting ETF risks and provides a novel framework for risk management and investment decision making, particularly in the Iranian financial market.

The present study aims to identify the most important variables affecting the fluctuations of gold prices. It is the first Iranian research in which the fluctuations in this market are modeled using non-linear Bayesian Model Averaging (BMA) and deep neural network approaches.

It is applied research where monthly data collected from 2010 to 2022 were used. It evaluates 35 factors playing a role in gold price fluctuations. GARCH and random fluctuation models are used to extract gold price fluctuations. TVPDMA, TVPDMS, and BMA models are used to identify the most important variables causing gold price fluctuations. Furthermore, the deep learning approach is used to investigate how effective the selected variables are in gold price fluctuations.

The results indicated that Support Vector (SV) models were more accurate than GARCH models in capturing fluctuations and that BMA outperformed TVPDMA and TVPDMS. Additionally, 12 variables were identified as influential in gold price fluctuations, with in-market factors playing a more significant role than out-of-market factors. The study also employed Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and Multi-Layer Perceptron (MLP) neural network models in deep learning mode to predict gold price fluctuations. It was concluded that global interest rates had the most significant impact on fluctuations in the gold price, with the Pivot Point DeMark's Index making the greatest contribution.

Originality/Value: 

The gold market is known for its volatility, and accurate predictions about its future can significantly impact decision-making. Understanding the gold price and making correct forecasts can help inform decisions about buying and selling gold in global markets, and determine the most favorable times for transactions and investments. Therefore, it is crucial to accurately predict the gold price from various perspectives. This research attempted to develop an intelligent model for forecasting fluctuations in the gold price.

Today, maintenance and repair have become very important in the manufacturing industry. An ecient solution to prevent downtime is to predict equipment failure. Therefore, accurate and correct prediction of breakdown events in the eld of predictive maintenance can be very useful. In general, each prediction will be accompanied by a certain amount of error, which in various ways tries to control this error or limit it to a reasonable amount. In this thesis, a framework has been proposed that species when the system under review will need maintenance and repairs to prevent downtime as much as possible. Therefore, the main purpose of this thesis is to design and implement an ecient combination structure to accurately predict failure events using both standard statistical standard models and machine learning in predictive maintenance. The literature review results indicate that the use of these methods in recent years has led to extensive advances in the eld of providing accurate forecasts and subsequently improved the level of decisions made by managers and decisionmakers. The proposed model is used to predict failure events in benchmark data related to the truck air pressure system. Finally, the performance of the proposed model is compared with other data-driven techniques individually and in combination, which includes logit models, support vector machines, and multilayer perceptron neural networks. According to the numerical values obtained from the nal analysis, the results indicate that the backup vector machine model has higher prediction accuracy than other single models, and also the results indicate the eciency and eectiveness of the proposed parallel combination structure compared to the use of models individually and in series combination in modeling and forecasting issues. The parallel hybrid model improved the accuracy of predictions by an average of 11% in test data and 7% in training data. Therefore, due to the greater accuracy in combining classical statistical models and machine learning in parallel, the use of this combined method to improve the accuracy of predictions in the eld of predictive maintenance is recommended for future studies.

The bullwhip effect in the supply chain could lead to fluctuations such as extra inventory and delayed order. In the meantime, proper demand forecasting can significantly resolve these fluctuations by eliminating the bullwhip effect. The present study considers forecasting of retailer demand, suppliers and calculates the bullwhip effect in the two-level supply chain. Markov switching model and autoregressive model along with moving average are used to predict retail demand and supplier, respectively. The results showed that the lot size policy-based ordering system can reduce or completely remove the bullwhip effect. Besides, the bullwhip effect is appeared in both levels of the supply chain during utilizing the rm(q0,Qm) order-point policy-based ordering system.

Corona is currently the world's health crisis and the biggest challenge humans have experienced since World War II. Given the epidemic of the disease, it is invaluable to forecasting the number of cases and the resulting deaths to better understand the current situation and provide a short-term plan by managers. Accordingly, in this paper, a neuro-fuzzy network model is proposed to forecast the number of cases and deaths in countries that are most affected by this disease. The performance of the proposed neuro-fuzzy network has been compared with time series forecasting neural network as well as radial basic functions neural networks. The proposed model is able to predict the number of cases and deaths from the disease for a period of the next 15 days at a lower error rate.