Data Mining for “Big Archives” Analysis: a Case Study

摘要

We present a case of archival analysis using a combinationof data mining methods. The team of researchers, composedby archivists and computer scientists, used a collection ofdeclassified Department of State Cables as a case study. Themethods implemented included Support Vector Machine(SVM) and Association Rule Mining. Combined in ananalysis workflow, the results of the different methodsallowed the team to identify the different security classes,understand how they changed over time and generatedescriptions for the cables in each class. The interpretationof results also allowed understanding contextual aspects ofthe collection. Until now, the use of data mining for archivalanalysis and processing has not been thoroughly exploredby the archival community. This study constitutes a seminalroadmap to understand how to apply, interpret and integratedata mining with the archivists experience and judgment incollaboration with computer scientists. It proposes aninductive approach to archives analysis and the possibilityof verifying processing decisions.