Beyond keyword counts: an information-weighted NLP index of corporate digitalization
Yang Guan, Shuo Wang, Jianxu Liu · Applied Economics Letters · 2025
This study proposes a Multidimensional Information-Weighted Text-Mining (MIW-TM) framework to measure the digitalization levels of Chinese A-share listed companies from 2006 to 2023, addressing two key limitations of traditional text-based methods: ‘nano-fake’ (superficial or vague digital descriptions that overstate true digitalization) and ‘abandoned truth’ (the loss of meaningful information when frequency-based weighting methods underestimate statistically rare but important terms). We find that the MIW-TM framework identifies leading digital enterprises more accurately, digital development has accelerated sharply since 2021, and substantial heterogeneity exists across both enterprise types and digital technologies.