RETRACTED ARTICLE: Neural dynamics: unraveling the impact of digital economy on regional growth

This research investigates the intricate dynamics between digital economy indicators and regional economic growth, leveraging artificial neural networks as a robust analytical framework. This study employs a multi-layer perceptron (MLP) with a specific architectural configuration to examine the intricate connections between digital economy indicators and regional economic growth. The MLP comprises an input, two hidden, and an output layer. The complexity of the relationships within the data set drove the decision to include two hidden layers. Multiple hidden layers allow the model to capture intricate patterns and non-linear dependencies, enhancing its capacity to discern the nuanced impact of digital economy indicators on regional economic dynamics. The number of neurons in each layer is tailored to optimize the models. It utilizes specific techniques in data cleaning, including log and Box–Cox transformations, to address skewed distributions and stabilize variance. Data splitting incorporates a systematic approach with continuous validation throughout the analysis, iteratively refining cleaning and pre-processing steps based on gained insights. Through rigorous training, validation, and evaluation processes, the MLP demonstrates robust operations, showcasing competitive metrics, such as an accuracy of 55%, precision of 34%, recall of 68%, and an F1 score of 45%. The anticipated outcome is a nuanced understanding of the complex relationships between the digital economy and regional economic dynamics, supported by a comprehensive ROC analysis. This model, with its carefully tuned architecture and specific inclusion of two hidden layers, outperforms three other studies in the same domain, contributing valuable insights to the discourse on the impact of the digital economy on regional growth.