Applying Text Mining in Operations and Engineering Management: Guidelines and Challenges in Sentiment Analysis and Topic Modeling
这篇综述回顾了2011至2025年间文本挖掘在运营与工程管理研究中的应用，提供了从数据收集到验证的完整指南，并重点讨论了情感分析和主题建模中的语义歧义与模型不稳定等挑战及改进方法。
Text mining helps researchers extract meaningful insights from large amounts of unstructured text. This study reviews how text mining has been used in operations management (OM) and engineering management (EM) research between 2011 and 2025, with a focus on articles published in leading journals in both fields. Our review shows that the use of text mining has increased rapidly in recent years, reflecting its growing importance as a research tool. A central contribution of this study is the development of clear and consistent guidelines for applying text mining in OM and EM research. These guidelines cover the full research process, including data collection, preprocessing, feature extraction and selection, modeling, interpretation, and validation. The review also examines two major challenges in current applications: sentiment analysis and topic modeling. In both areas, researchers often face problems such as semantic ambiguity and model instability, which can reduce the reliability and interpretability of findings. To address these issues, we review and recommend approaches such as Stable Latent Dirichlet Allocation and aspect-based sentiment analysis, drawing on established literature to identify methods that can improve analytical robustness. Overall, this review offers both researchers and practitioners a practical resource for applying text mining more effectively, consistently, and with greater confidence in OM and EM contexts.