Using Case-Based Reasoning to Support the Indexing and Retrieval of Incident Reports

Incident reporting systems can be used to detect problems before they result in an accident. They can also be used to strengthen the defences that lead to the detection and resolution of potential problems. There are also significant limitations. For instance, it is difficult to support long-term participation from all elements of a workforce. In spite of these difficulties, incident reporting schemes are increasingly being introduced into many industries. This growth is creating new challenges. In particular, it can be difficult to spot emerging trends and common features amongst the thousands of reports that are submitted to many international schemes. Traditional databases offer little support here because query formation often defeats even relatively skilled analysts. Similarly, free-text search engines have technical limitations that may it difficult to identify incidents in which certain causes were NOT a factor. As a result many of these systems yield results that have poor precision and low recall values. This paper, therefore, argues that alternative techniques must be developed to support the indexing and retrieval of similar cases from within the growing body of evidence in large-scale incident reporting schemes. In particular, we show how case-based reasoning techniques can be extended from the domain of decision support to help analysts retrieve information about previous incidents. The US Navy’s Conversational Decision Aids Environment (NaCoDAE) is used to illustrate this argument. In particular, we have applied it to two datasets from the US Aviation Safety Reporting System (ASRS). This initial work has exploited two different classification schemes. The first was based on part of the FAA’s ASRS reporting process. The second was based on the more general Eindhoven classification method.