Unlocking machine learning for social sciences: The case for identifying Industry 4.0 adoption across business restructuring events.

In recent years, advancements in machine learning (ML) have facilitated the utilisation of big data across various academic disciplines. Nonetheless, these techniques still require a high-level of programming and data science expertise, making them inaccessible to many researchers and hindering the potential for knowledge advancements. This paper presents a framework for identifying the adoption of Industry 4.0 (I4.0) technologies among European firms that have undergone restructuring events. Existing studies on I4.0 adoption rely on diverse data sources at different levels of aggregation (e.g., countries, sectors, firms), spanning various time periods and technological domains. While this diversity often complicates result comparison, it also drives researchers and institutions to explore new data sources to assess technology adoption. Our identification methodology is based on the implementation of ML techniques using STATA, a well-established and user-friendly statistical software. We offer a step-by-step guide based on recently developed commands, allowing for comparison of model performance and analysis of model features. Our findings underscore the potential of ML algorithms as a robust tool for collecting new firm-level data on I4.0 adoption. Specifically, we observe that business restructuring events predicted as I4.0-related conform to adoption patterns identified in prior studies, across countries, sectors and over time. • This study presents machine learning models to predict Industry 4.0 adoption. • A methodology is presented to explore the application of data science techniques in social sciences. • The machine learning implementations are based on a set of recent commands for STATA. • Eurofound's data on 23,618 restructuring events were used to predict technology adoption from textual descriptions. • Predicted adoption patterns conform to those observed across European countries, sectors, and over time. [ABSTRACT FROM AUTHOR]