Abstract
From patient waiting time to consumer shopping, firms collect more and more human behavior data to assist their decision making process. This trend in business also affects academic research, especially in operations management (OM), a research area that often relies on mathematical modeling to guide business decisions. However, it is both time and labor intensive to identify applications and opportunities that use behavioral big data (BBD) in the large and growing published literature. In this paper, we introduce a procedure that applies various data mining approaches to survey a vast number of research articles across three different OM journals, and identify articles that use BBD. The goal is to reduce the number of articles that must be read manually and yet reduce the false negatives (missed BBD papers); in other words, in this classification task we emphasize the importance of sensitivity over specificity with respect to detecting BBD papers. Testing different feature engineering and classification approaches, we find that the highest sensitivity and specificity are provided by a Random Forest classifier, applied to a bag-of-words set of features.