Journal of Engineering, Project, and Production Management, 2026, 16(5), 2026-0033
HNN-DIC: Spatio-Temporal Evaluation of SME Digital Innovation Capabilities Using Industrial Internet Platform Data
1
Postgraduate Student, Business Administration, Inje University, Gimhae,
South Gyeongsang Province, South Korea
2 Postgraduate Student, Business
Administration, Dong-eui University, Busanjin-gu, Busan, South Korea
3 Ph.D., School of Civil
Engineering and Architecture, Guangxi Minzu University, China, E-mail:
lijing-kang@outlook.com (corresponding author).
Project Management
Received April 28, 2026; revised June 4, 2026; accepted July 3, 2026
Available online July 12, 2026
Abstract: In the context of the digital economy, scientifically assessing the Digital Innovation Capability (DIC) of Small and Medium-Sized Enterprises (SMEs) is essential for achieving sustainable competitive advantages. Traditional evaluation approaches often struggle to process multi-source, high-dimensional and nonlinear digital data and lack dynamic assessment capability and model interpretability. To address these limitations, this study proposes a data-driven intelligent evaluation framework that combines industrial internet platform data with deep learning techniques to enable dynamic, interpretable SME-DIC assessment. Specifically, a Hybrid Neural Network-based Digital Innovation Capability evaluation model (HNN-DIC) is developed by integrating a One-Dimensional Convolutional Neural Network (1D-CNN) and a Long Short-Term Memory (LSTM) network. The proposed framework simultaneously captures local spatial correlations and long-term temporal evolution patterns among digital innovation indicators. Based on DIC, a multidimensional evaluation system containing nine core indicators is constructed across four dimensions: technological embedding, dynamic adaptation, resource coordination, and innovation output. The proposed model is trained and validated using two years of panel data collected from 500 manufacturing SMEs operating on an industrial internet platform. Experimental results demonstrate that HNN-DIC achieves a coefficient of determination (R²) of 0.931 on the test set, with a Mean Absolute Error (MAE) of 1.72 and a Root Mean Square Error (RMSE) of 2.28, significantly outperforming benchmark models such as Random Forest and XGBoost. Furthermore, SHapley Additive exPlanations (SHAP) analysis identifies “data analysis activity” and “external collaboration breadth” as the most influential drivers of SMEs digital innovation capability. The study contributes a domain-specific hybrid intelligent assessment framework for SMEs-DIC evaluation and provides an interpretable decision-support tool that can be deployed in cloud-based digital platform ecosystems to support enterprise self-diagnosis, strategic decision-making, and policy formulation.
Keywords: Digital innovation capability; small and medium-sized enterprises; digital platform; hybrid neural network.
Copyright © Journal of Engineering, Project, and Production Management (EPPM-Journal).
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
Requests for reprints and permissions at eppm.journal@gmail.com.
Citation: Guo, M., Kai, Q., and Kang, L. (2026). HNN-DIC: Spatio-Temporal Evaluation of SME Digital Innovation Capabilities Using Industrial Internet Platform Data. Journal of Engineering, Project, and Production Management, 16(5), 2026-0033.
