Enhancing economic competitiveness analysis through machine learning: Exploring complex urban features.

Area of Science:

• Urban Studies
• Economic Geography
• Computational Social Science

Background:

• Urban economic competitiveness is key to development and understanding regional disparities.
• Traditional regression models struggle with complex, nonlinear relationships among urban features.
• Cities are complex systems, requiring advanced methods beyond narrow feature analysis.

Purpose of the Study:

• To develop a novel analytical model for urban economic competitiveness using deep learning.
• To accurately classify urban economic competitiveness by capturing intricate feature interrelationships.
• To address the challenge of limited sample sizes in urban deep learning research.

Main Methods:

• Constructed a dataset of 1008 features from 283 Chinese prefecture-level cities.
• Employed a convolutional neural network (CNN) for feature interrelationship analysis and classification.
• Utilized deep convolutional generative adversarial networks (DCGANs) for data augmentation to enhance model performance.

Main Results:

• Developed a precise and stable analytical model for urban economic competitiveness classification.
• Demonstrated that data augmentation with DCGANs significantly improved the CNN model's accuracy and generalization.
• Successfully captured complex, nonlinear interrelationships among a large set of urban features.

Conclusions:

• Deep learning offers a powerful approach to studying urban economic competitiveness and disparities.
• The developed CNN-DCGAN model provides a robust solution for analyzing complex urban systems with limited data.
• This research offers valuable insights into regional development disparities and a methodological advancement for urban big data analysis.