Regional Economic Prediction Model Using Backpropagation Integrated with Bayesian Vector Neural Network in Big Data...

Area of Science:

• Artificial Intelligence
• Computational Economics
• Machine Learning

Background:

• Accurate economic growth forecasting is essential for effective national economic development policies.
• Artificial Neural Networks (ANNs) are powerful tools for modeling complex functions and analyzing real-world data.
• Existing methods require improvement in speed and accuracy for regional economic predictions.

Purpose of the Study:

• To propose a regional economic prediction model utilizing Artificial Neural Networks (ANNs).
• To enhance the training speed and prediction accuracy of economic forecasting models.
• To provide a viable and accurate model for practical application in economic policy formulation.

Main Methods:

• Integration of Bayesian vector neural network (BVNN) with the backpropagation (BP) model.
• Knowledge-based computer analysis using neural networks for data extraction and classification.
• Feature selection techniques including discretization, reduction, and importance ranking to optimize NN structure.

Main Results:

• The proposed model achieved high accuracy across multiple datasets: 98% for WEO, 92% for APDREO, and 98% for AFRREO.
• Significant improvements in training speed and prediction accuracy were observed compared to existing methods like FWA-SVR and LSTM.
• The model demonstrated effectiveness in tackling nonlinear economic problems, achieving up to 97% GDP prediction accuracy.

Conclusions:

• The developed neural network-based model is a successful and viable tool for regional economic forecasting.
• The integration of BVNN and BP, coupled with optimized feature selection, significantly enhances model performance.
• The model offers practical value for policymakers by providing accurate and efficient economic growth predictions.