Title: Unsupervised Learning Techniques for Analyzing Firm-Level Determinants of Total Factor Productivity

Abstract: This study employs unsupervised learning methods—specifically principal component analysis, self-organizing maps, and clustering—to identify the key characteristics that drive firm-level total factor productivity. By adopting a bottom-up perspective, the research effectively addresses the challenges posed by firm heterogeneity and offers fresh insights into traditional industry classifications. Leveraging the extensive data available in the ORBIS database, the analysis spans the pre-pandemic era (2015–2019) and the immediate post-Covid period (2020). The findings indicate that, across both timeframes, the primary drivers of productivity growth are linked to profitability, credit and debt metrics, cost and capital efficiency, and the effectiveness of firms’ R&D efforts. Furthermore, the study establishes a linear correlation between these determinants and productivity growth.

Source: arXiv Generated at: 2026-06-02 00:00:00 UTC