Cause analysis of new incidents by using failure knowledge database

Abstract

Root cause analysis of failed projects and incidents is an important and necessary step to working out measures for preventing their recurrences. In this paper, to better analyze the causes of failed projects and incidents, we propose a novel topic-document-cause(TDC) model that reveals the corresponding relationships among topics, documents, and causes. We use the JST failure knowledge base to construct a TDC model with machine learning methods such as LDA and perceptron. The experimental results show that our approach performed better at discovering the causes of failures for projects and incidents. © 2012 Springer-Verlag.