Regression and Machine Learning Methods to Predict Discrete Outcomes in Accounting Research
本文为会计研究者提供了用逻辑回归和机器学习模型预测离散结果的实用指南，并以预测美国证券交易委员会调查为例，强调样本外准确性和结果呈现的要点。
ABSTRACT Predictive modeling focuses on iteratively trying various combinations and transformations of a set of variables to generate a decision rule that predicts outcomes for new observations. Although accounting researchers have demonstrated interest in predictive modeling, we identify a lack of accessible and applied guidance on this topic for accounting settings. This issue has become more salient with the increasing availability of machine learning models that use unfamiliar terminology, are estimated using algorithms, and produce different outputs than other models used for causal inference. To overcome this gap, we provide an overview of how to predict discrete outcomes with logistic regression and machine learning models used in recent studies. We also include guidance and a comprehensive example—predicting investigations by the U.S. Securities and Exchange Commission—that illustrates the elements of the prediction process, highlighting the importance of out-of-sample accuracy and unique aspects in the presentation of a prediction model's results. Data Availability: The data and code to replicate our example are available upon request. First, those interested must request and receive the SEC investigation data from Blackburne et al. (2021). Next, we will provide code to merge the SEC investigation data with Compustat and CRSP and replicate our analyses. JEL Classifications: C10; C25; C45; C53; M48.