The Aviation Safety Data Mining Workbench offers three key data mining techniques designed for aviation safety analysis. The first technique, FindSimilar, employs both information retrieval and data mining methods to examine text and structured data. Next is FindAssociations, which delves into the collection of incidents to uncover subsets that exhibit interesting correlations. Finally, FindDistributions hones in on a specific field or attribute of the incidents. This method assesses the overall distribution for that field, extracting subsets of data to identify how each subset’s distribution compares to the overall one. Subsets that show significant differences from the main distribution are flagged as particularly intriguing. This technique is essential for spotting anomalies that could warrant further investigation. The workbench was developed by MITRE.
