How to conduct efficient and objective literature reviews using natural language processing: A step‐by‐step guide for marketing researchers
本文为营销研究者提供了一份分步指南，说明如何利用自然语言处理（包括TF-IDF和LDA主题建模）来高效、客观地进行文献综述，并附有可直接使用的Python代码。
Abstract Literature reviews are crucial for attaining a full understanding of the key topics and latest trends in research and instrumental in identifying important research gaps. Unfortunately, conducting literature reviews can be time‐consuming, and the outcomes are frequently subjective. Hence, to address such limitations, we detail an alternative, recent approach to conducting literature reviews. In this research, we outline the steps involved in conducting a literature review via natural language processing. Specifically, we illustrate how to (1) select relevant papers using term frequency‐inverse document frequency and (2) perform topic modeling analysis through latent Dirichlet allocation to identify key research topics. This study and the associated ready‐to‐use Python code provide researchers, including those in consumer behavior, with detailed guidance on how to use natural language processing in their literature reviews.