Machine Learning Approaches to Technological Innovation and Corporate Financialization

This study develops a comprehensive machine learning framework to explore the intricate relationship between corporate financialization and technological innovation using firm-level data from 2010 to 2024. The dataset integrates financial indicators, innovation metrics, and contextual variables such as firm size, industry sector, and geographic location, sourced from Compustat, PATSTAT, and the USPTO. Continuous variables are standardized using z-score normalization, and feature engineering techniques compute key financial ratios to capture proportional relationships. Random Forest Regression (RFR) models innovation outcomes based on financial data, with performance evaluated by R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>, MSE, and MAE metrics. K-means clustering, optimized via the elbow method, classifies firms into typologies reflecting financialization and innovation profiles. Additionally, term frequency-inverse document frequency (TF-IDF) analysis extracts innovation sentiment scores from textual financial records, which enhance both regression and clustering models. The integrated framework provides robust empirical insights and strategic profiling of how financial strategies influence innovation development across diverse firm types and economic cycles.