|
Designing an Intelligent System for Inflation Forecasting and Management in Iran Using Artificial Neural Networks and Fuzzy Goal Programming
|
Reza Najafi1   , Nazi Mohammadzadeh Asl *2 , Shahriyar Nessabian1 , Marjan Daman keshideh1 |
1- Department of Economics, CT.C., Islamic Azad University, Tehran, Iran
2- Department of Economics, CT.C., Islamic Azad University, Tehran, Iran , n.asl34@iau.ac.ir |
|
|
Abstract: (5 Views) |
This paper presents an intelligent framework based on Artificial Neural Networks (ANN) and Fuzzy Goal Programming (FGP) to analyze and forecast inflation in Iran. Time-series data spanning from 1959 to 2023 (1338–1402 SH), encompassing key macroeconomic variables such as exchange rates, liquidity, gold prices, oil prices, Gross Domestic Product (GDP), and unemployment rates, are utilized. The findings indicate that the causal structure of inflation underwent significant structural shifts across two distinct historical periods: pre- and post-Islamic Revolution. In the pre-revolutionary period, endogenous variables—predominantly liquidity growth and inflationary expectations—played a dominant role. Conversely, in the post-revolutionary period, exchange rate fluctuations, international sanctions, and subsidy policies exerted a more pronounced impact on driving inflation. Furthermore, integrating the ANN model with FGP improved forecasting precision, reducing the Mean Absolute Error (MAE) to below 3% under standard economic conditions and 5% during crisis scenarios. Policy scenario analyses—focusing on monetary discipline, sanctions relief, and subsidy reforms—demonstrate that this hybrid computational approach significantly enhances policy decision-making in managing inflation dynamics.
|
Article number: 5 |
| Keywords: Inflation, Artificial Neural Network (ANN), Fuzzy Goal Programming (FGP), Sanctions, Economic Policies, Intelligent Forecasting. |
|
|
Full-Text [PDF 624 kb]
(3 Downloads)
|
Article type: Research |
Subject:
General
Received: 2026/04/30 | Accepted: 2026/08/5
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
