Regional economic forecast using Elman neural networks with wavelet function

Area of Science:

• Economics
• Artificial Intelligence
• Time Series Analysis

Background:

• Guangdong province leads China's economy, necessitating accurate Gross Domestic Product (GDP) prediction to ensure sustained high-quality growth.
• Existing forecasting models may lack the precision required for dynamic regional economic analysis.

Purpose of the Study:

• To develop and validate a novel forecasting model for Guangdong's regional economy.
• To enhance the accuracy and efficiency of economic prediction using advanced computational techniques.

Main Methods:

• Implementation of an Elman neural network integrated with a wavelet function for GDP forecasting.
• Comparative analysis against existing models to evaluate forecast accuracy and precision.
• Assessment of the model's scalability and performance with large datasets.

Main Results:

• The proposed Elman neural network with wavelet function achieved a forecast accuracy of 0.971.
• The model demonstrated superior precision and lower errors compared to competing methods.
• Investment in education was identified as a significant positive driver of regional economic development.

Conclusions:

• The wavelet-enhanced Elman neural network is highly effective for regional economic forecasting, adaptable to various scenarios.
• The model offers improved forecast accuracy and training efficiency, particularly for large-scale datasets.
• Wavelet functions provide a cost-effective method to enhance neural network performance without increasing architectural complexity.