Scholarlink

<p>As Europe advances toward the vision of Industry 5.0 (where technological progress is expected to reinforce sustainability, resilience, and human-centric development) understanding the uneven geography of digital transformation becomes increasingly critical. This study proposes a novel analytical framework to examine regional disparities in digitalisation across EU NUTS-2 (European Union Nomenclature of Territorial Units for Statistics) regions between 2012 and 2023. We derive a data-driven representation of digitalisation by combining an autoencoder architecture for latent-feature extraction with deep embedded clustering to uncover structural patterns of technological readiness. To enhance transparency, the latent dimensions are interpreted ex post using explainable-AI techniques (SHAP—SHapley Additive exPlanations), which show that human-capital endowment, R&amp;D (research and development) intensity, high-technology employment, and tertiary education show the strongest overall impact on regional digital advancement, as confirmed by the SHAP importance results (Fig. <InternalRef RefID="Fig19">19</InternalRef>) and structural equation modelling (SEM) loading patterns (Table <InternalRef RefID="Tab6">6</InternalRef>). The resulting clusters reveal a persistent tripartite stratification, suggesting that digitalisation has not yet facilitated broad convergence; instead, it reinforces existing divides by favouring regions with strong knowledge and innovation infrastructures. These findings highlight the need for targeted public policies centred on skills formation, research capacity, and inclusive infrastructure in order to ensure a just and cohesive transition toward a knowledge-based economy. Beyond its empirical contribution, the study provides a replicable methodological pathway for monitoring digital progress and supporting place-based policy design</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Regional disparities in the transition to industry 5.0: an algorithmic approach to digitization

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月2日

查看全文
  • Monica Laura Zlati,
  • Camelia Madalina Beldiman,
  • Iulian Vasiliev,
  • Valentin-Marian Antohi,
  • Costinela Fortea

摘要

As Europe advances toward the vision of Industry 5.0 (where technological progress is expected to reinforce sustainability, resilience, and human-centric development) understanding the uneven geography of digital transformation becomes increasingly critical. This study proposes a novel analytical framework to examine regional disparities in digitalisation across EU NUTS-2 (European Union Nomenclature of Territorial Units for Statistics) regions between 2012 and 2023. We derive a data-driven representation of digitalisation by combining an autoencoder architecture for latent-feature extraction with deep embedded clustering to uncover structural patterns of technological readiness. To enhance transparency, the latent dimensions are interpreted ex post using explainable-AI techniques (SHAP—SHapley Additive exPlanations), which show that human-capital endowment, R&D (research and development) intensity, high-technology employment, and tertiary education show the strongest overall impact on regional digital advancement, as confirmed by the SHAP importance results (Fig. 19) and structural equation modelling (SEM) loading patterns (Table 6). The resulting clusters reveal a persistent tripartite stratification, suggesting that digitalisation has not yet facilitated broad convergence; instead, it reinforces existing divides by favouring regions with strong knowledge and innovation infrastructures. These findings highlight the need for targeted public policies centred on skills formation, research capacity, and inclusive infrastructure in order to ensure a just and cohesive transition toward a knowledge-based economy. Beyond its empirical contribution, the study provides a replicable methodological pathway for monitoring digital progress and supporting place-based policy design