Can Machine Learning Predict ICT Business-Process Innovation?
Taeung Eom (Pukyong National University), Chungwon Woo & Dongphil Chun (Pukyong National University) Technology Analysis & Strategic Management, 36(9), 2271–2283, 2024 · DOI: 10.1080/09537325.2022.2132927
TL;DR — Using the 2020 Korea Innovation Survey, machine learning predicts ICT business-process innovation (a core form of digital transformation). The random forest model performed best, and information sources were the single most critical predictor — firms that tap external information are the ones that convert ICT into process innovation.
Key findings
1. Innovation forecasting works. ML models — random forest in particular — accurately predict which firms achieve ICT business-process innovation, despite the high uncertainty of fast-evolving ICT.
2. External information beats internal resources as a predictor. Among variable groups, access to information sources mattered most for predicting innovation performance.
3. Managerial takeaway. Digital transformation programs should institutionalize external scanning (customers, suppliers, universities, competitors) rather than rely on internal IT capability alone.
국문 요약
2020 한국기업혁신조사(KIS) 자료에 기계학습을 적용해 ICT 비즈니스 프로세스 혁신(디지털 전환) 성과를 예측. 랜덤포레스트 모형의 예측력이 가장 높았고, 설명변수 중 정보원천(외부 정보 활용)이 가장 중요한 예측 요인이었다. 디지털 전환을 추진하는 기업은 외부 정보를 적극 활용해야 한다는 시사점.
FAQ
Q. Why use machine learning instead of regression here? ICT evolves faster than other technologies and innovation outcomes are highly uncertain and non-linear; tree-based ensembles capture these interactions and rank variable importance directly.
Q. Related work by the authors? Jang, Ko, Chung & Woo (2023, TASM) — also on the Korean Innovation Survey — shows how external collaboration configurations drive product vs. process innovation.
How to cite
APA: Eom, T., Woo, C., & Chun, D. (2024). Predicting an ICT business process innovation as a digital transformation with machine learning techniques. Technology Analysis & Strategic Management, 36(9), 2271–2283. https://doi.org/10.1080/09537325.2022.2132927
BibTeX:
@article{eom2024ict,
title={Predicting an ICT business process innovation as a digital transformation with machine learning techniques},
author={Eom, Taeung and Woo, Chungwon and Chun, Dongphil},
journal={Technology Analysis \& Strategic Management},
volume={36}, number={9}, pages={2271--2283}, year={2024},
doi={10.1080/09537325.2022.2132927}}