محمدرضا اصغری اسکویی

One of the key issues in financial stability studies and prudential policies is systemic liquidity risk, in other words, the risk of liquidity problems at the same time and in several financial institutions. The fact that in the complex network of interbank market interconnections, the lack of liquidity is provided by spreading it between financial institutions, can lead to "systemic contagion" in the banking network. This research aims to investigate the existence of systemic contagion in the banking network of Iran and the effect of the Basel Committee's prudential regulations on interbank contagion based on the balance sheet data of 25 banks that are members of the interbank market in the years 2006-2018. For this purpose, based on the agent-based approach, Iran's banking network has been modeled and simulated. In this model, banks and the central bank are intelligent agents. Based on the obtained results, prudential regulatory requirements change banks' adaptive strategies, and banks put increasing the capital adequacy ratio on their agenda. In addition, this research shows that although the guidelines of the Basel Committee have succeeded in reducing contagion in the long run and making the banking network more stable, banks also reduce the supply of loans to the real sector, which can lead to a crunch in the economy.

agent-based modeling is an emerging computational technique that makes it possible to simulate complex economic systems, including the banking network, with a bottom-up approach. In this paper, the country's banking network is simulated with an intelligent multi-agent modeling model and indicates that these agents behave based on the adaptive learning. This modeling has been done with the aim of examining and evaluating the impact of regulatory policies on the interbank market and based on the balance sheet data of  25 member banks of the interbank market in the years 2006-2019. To assess the impact of regulatory policy, the scenario of having Clearing House to reduce non-payments in the interbank market has been examined. Due to the learning of the agents, in this simulation, the results show the direct and indirect impact of regulations on the interbank market by changing the adaptive strategies of the agents. According to the results of this study, monitoring the interbank market through the Clearing House, solves the problem of information asymmetry in the interbank market and thus reduces financial contagion and increases the stability of the system.

Risk classification of insurance customers, based on the observable characteristics, can significantly help insurers mitigate losses, classify their customers and prevent adverse selection. This paper aims to study losses occurred in motor Third Party Liability (TPL) insurance and predict customers’ risk of loss.

With the help of four supervised algorithms namely; decision tree, SVM, naïve Bayes and neural network hidden pattern of data is discovered to classify customers of TPL insurance. Furthermore, the imbalanced dataset was the main challenge for implementing machine learning and data mining techniques which will be discussed throughout the article.

The dataset contains more than 400,000 observations for five years from an Iranian insurance company. It also has five variables of which four are independent: car type, car group, plate type, car age; and one binary dependent variable: financial loss. Comparing the model performances, decision tree is the most efficient (F1=0.72±1).

The model provides prioritization of independent features as follows: car type, plate type, car age, car group. Findings also suggest that to obtain more accurate prediction on claims and high-risk customers, more features concerning drivers’ traits are required. 

I International relations are full of complexities due to their multifaceted and multilateral nature. To understand decision making processes and the payoffs of their strategies, players are enabled to utilize their capabilities to impact the strategic decision payoffs. As an example of this decision structure we can refer to the international disputes and conflicts including the sanctions. The Islamic Republic of Iran’s interests in international arena have been influenced by the changing interests of the other stakeholders in the coalition of the United States sanction against Iran. In order to analyze and predict the equilibriums of the players positions in the complex international space of the sanctions, game theory, mechanism design, multi-agent systems, and artificial intelligent as new instruments of decision theory are utilized to resolve the transactions and processes influenced by the human decisions. In this research, a policy spectrum is defined according to the strategies of Iran, the United States of America and other stakeholder’s countries. Modeling and simulating the behavior of players on this continuous spectrum which includes two extremes of capitulation and confrontation shows that in the current situation, Iran’s equilibrium and dominated strategy is emphasizing the maintenance of the current agreement in JCPOA, although this position is in the lower range of the spectrum of the current agreement. In these sanctions there is a possibility of a tragedy of the commons. In order to prevent its occurrence, the results of this study imply that the dominant strategy, given that all the players are rational, is to preserve stability, security, and integrity of Iran as a regional power. In order to achieve a stable equilibrium of the game, the stakeholders try to maximize the social welfare function instead of individual participants’ payoff. The rational strategy of Iran is to stay in JCPOA, strengthening political ties with the European players, and to empower its own military and social securities.

We develop an agent-based model to study the effects of the electricity market reform on the electricity prices, the power plants’ technology mix as well as their capacity utilization and profits. The reform’s main objective is to open the electricity market to more competition and increase the efficiency. The new wholesale electricity market will operate based on the one-day-ahead auctions organized by the independent system operator. We set up two scenarios in which the market clearing mechanism changes from the pay-as-bid to market clearing price (MCP) and the fuel subsidy to power plants will be removed. The effects of the two scenarios will be analyzed on the market prices, plant revenues, market shares, and power generating units’ capacities. The simulation results show that the electricity prices are higher during the peak load mainly due to the entry of the higher cost plants during those periods. Electricity prices are lower under the MCP scenario and higher under the fuel subsidy removal scenario, leading to lower and higher revenues for the power plants, respectively. The results also indicate that the market shares and the capacity utilization of the more efficient plants will increase under both scenarios.

With increasing competition in the wholesale Electricity markets and advances in behavioral economics in recent years, the multi-agent modeling approach has been applied widely to simulate the outcome of the markets. The electricity market consists of power generating agents that compete over production in daily auction conducted by an independent system operator (ISO). The market clearing mechanism can be seen as a static game that repeats every hour. In this game, an agent proposes her price for the next day and the ISO chooses the best proposals that minimizes the total costs given the demand and the technical constraints. Agents are also assumed to learn from the outcomes and adjust their biding strategy accordingly. In this paper, we develop an agent-based model for the day-ahead and pay-as-bid electricity market in Iran. The objective is to compare the outcome of the market measured by the agents profit and the time to converge using three different strategies: greedy, random and reinforcement learning. The simulation results indicate that the reinforcement learning leads to higher profits with a faster convergence rate than the other two strategies.

Given the overwhelming amount of information on the web, users face many options when selecting products or services. Recommender systems build a model based on information from user's past choices and ratings, related or trusted individuals, previously selected products, and the features of such products; the system then prioritize items to recommend them to the user based on this model. Trust aware method will use the trust network between users for estimating products ratings. Researchers have been interested in subject of trust in different facets because of different level of trust in professional fields. This article presents multi-faceted trust model for estimating product ratings, in which users and items are considered due to amount of dependency to each facet and also level of trust in it. Epinions dataset analysis indicates that distance dispersion of users’ choice in a multi-facted trust network is significantly lower than their distribution in a general trust network. Then baseline and similarity base models’ performance have been checked and compared in forms of general and multi-faceted. Model evaluation has been done based on Root Mean Squared Error and Epinions dataset separation in two groups of test and train and also croos validation method. Results indicate that estimation error has been averagely decreased 20% and improve recommender system performance obviously by considering trust component in multi-facted form.

The first step of forecasting time series is to build an appropriate model. Determining orders and estimation of ARMA model parameters is a challenging field in traditional statistical and intelligent methods. In this paper, genetic algorithm is used for parameter estimation and heuristic rules are used to determine orders of ARMA model. Heuristic rules are extracted according to time series properties. The data are selected using sliding time window. Model identification is carried out by using Bayesian information criterion (BIC). The mean squares error and the mean absolute percentage error are used to evaluate the results of prediction. The proposed method was applied to eight time series in different types, and the results were compared with results of statistical methods. The achieved result shows equivalent or superior performance for the proposed method in comparison with the classic method.

This paper proposes application of sliding window technique to time-delay neural network (TDNN) for prediction of financial time series. Neural network is a data-driven approach, in which we have huge data samples but limited information about the model structure. In this paper, we measure performance of the prediction and apply sliding window technique to select the most favorable neural network structure, time-delay taps and the most desirable training data size that result in the best prediction performance. The method was evaluated by using real data of share price of four firms traded in London Stock Exchange. The results show remarkable decrease for the root mean squared error, mean absolute percentage error and the linear regression of TDNN output offset.