/Macroeconomic-Variable-Demand-Forecasting-Using-Neural-Network-Prototype

Macroeconomic Variable Demand Forecasting Using Neural Network

Primary LanguageJupyter Notebook

Macroeconomic-Variable-Demand-Forecasting-Using-Neural-Network

This repository contains code for implementation of a research paper published in Journal of Mathematics and Statistics Studies. The article is available at the following link:

https://al-kindipublisher.com/index.php/jmss/article/view/5598

Abstract

In the competitive global corporate environment, accurate retail demand forecasting is crucial. Traditional forecasting methods often rely solely on micro variables, neglecting macroeconomic conditions that affect household demand for retail products. This study enhances forecasting models by incorporating external macroeconomic variables such as the Consumer Price Index (CPI), Consumer Sentiment Index (ICS), and the unemployment rate alongside time series data of retail sales.

Key Points

Macroeconomic Variables Used:

  • Consumer Price Index (CPI)
  • Consumer Sentiment Index (ICS)
  • Unemployment Rate

Methodology:

  • Train a Long Short-Term Memory (LSTM) model using both time series sales data and macroeconomic variables.

Findings:

  • The inclusion of macroeconomic variables improves the explanatory power of the model.
  • The LSTM model incorporating macroeconomic data outperforms models that do not include this information.
  • Demonstrates strong potential for industry application with improved forecasting capability.