An organization’s human resource management practices are essential for its competitive advantage. This study specifically examined human resource (HR) practices that predict corporate performance (employee turnover and firm sales) based on a backpropagation neural network (BPN)-based causality analysis. This study aims to test how to optimize human resource practices to improve organizational performance. This study elucidated the effect of HR practices and organizational-level factors on predicting employee turnover and firm sales. The BPN-based causality analysis revealed the relative importance of explanatory variables on firm performance. To test the model, it employed the Human Capital Corporate Panel open data on Korean companies’ HR practices and other characteristics. The analysis identifies causal relationships between specific HR practices and firm performance. The results show that compensation-related HR practices are most influential in predicting firm sales and employee turnover. Moreover, training-related HR practices were modest, and talent acquisition and performance management practices had relatively weak effects on the two outcomes. The study provides insights into how human resource practices can be optimized to improve firm performance and enhance organizational effectiveness. The findings of this study contribute to the growing body of research on the use of machine learning in HR management and suggest practical implications for managers’ insights to optimize HR practices.

中文翻译：

---

基于机器学习的人力资源实践对公司绩效的因果关系分析

组织的人力资源管理实践对其竞争优势至关重要。这项研究专门研究了基于反向传播神经网络 (BPN) 的因果关系分析来预测公司绩效（员工流动率和公司销售额）的人力资源 (HR) 实践。本研究旨在测试如何优化人力资源实践以提高组织绩效。这项研究阐明了人力资源实践和组织层面因素对预测员工流动率和公司销售额的影响。基于 BPN 的因果分析揭示了解释变量对企业绩效的相对重要性。为了测试该模型，它采用了人力资本企业小组关于韩国公司人力资源实践和其他特征的开放数据。该分析确定了特定人力资源实践与公司绩效之间的因果关系。结果表明，与薪酬相关的人力资源实践对预测公司销售额和员工流动率影响最大。此外，与培训相关的人力资源实践是适度的，人才获取和绩效管理实践对这两个结果的影响相对较弱。该研究提供了关于如何优化人力资源实践以提高公司绩效和提高组织效率的见解。这项研究的结果有助于越来越多的关于在人力资源管理中使用机器学习的研究，并为管理者优化人力资源实践的见解提出了实际意义。