A TECHNICAL FRAMEWORK FOR DATA-DRIVEN INDUSTRIAL TRANSFORMATION IN NORTHEAST CHINA

Abstract

This study proposes a comprehensive data-driven technical framework to support industrial transformation in Northeast China, using the Shenyang-Fushun Reform and Innovation Demonstration Zone as a case study. Based on real-world industrial value-added data spanning from 2000 to 2024, we develop and validate predictive models using SARIMA and LSTM neural networks to forecast regional industrial development trends. The framework integrates time series analysis, comparative regional analysis, and policy impact assessment to provide actionable insights for decision-makers. Our empirical analysis of monthly industrial value-added growth rates reveals significant seasonal patterns and structural changes over the 24-year period. The SARIMA model demonstrates superior performance in capturing seasonal variations, while the LSTM model shows enhanced accuracy for non-linear trend prediction. Comparative analysis with Beijing, Shanghai, Jiangsu, and Guangdong provinces reveals the unique characteristics of Northeast China's industrial development trajectory. The study provides evidence-based policy recommendations for accelerating industrial transformation through data-driven approaches, offering a replicable framework for similar post-industrial regions.