Machine Learning for Financial Statement Fraud

Abstract: Financial statement fraud (FSF) is a global concern representing a significant threat to financial system stability. In recent years, data-informed quantitative models have been developed to automate and reduce the manual auditing processes. Although the existing techniques have improved the detection rate of FSF, these are very limited and can be improved in terms of data, accuracy and efficiency, leading to more targeted and effective examinations. The main objective of this study is to develop four machine learning methods – Discriminant Analysis, Logistic Regression, Decision Trees and AdaBoost – in order to differentiate between fraud and non-fraud cases by assessing the likelihood of FSF using publicly available financial statement information.

Posted on Jan 12, 2016 in Seminario Aprendizaje de Máquinas, Seminars