Sequential pattern mining of multimodal streams in the humanities

Abstract

Research in the humanities is increasingly attracted by data mining and data management techniques in order to efficiently deal with complex scientific corpora. Particularly, the exploration of hidden patterns within different types of data streams arising from psycholinguistic experiments is of growing interest in the area of translation process research. In order to support psycholinguistic experts in quantitatively discovering the non-self-explanatory behavior of the data, we propose the e-cosmos miner framework for mining, generating and visualizing sequential patterns hidden within multimodal streaming data. The introduced MSS-BE algorithm, based on the PrefixSpan method, searches for sequential patterns within multiple streaming inputs arriving from eye tracking and keystroke logging data recorded during translation tasks. The e-cosmos miner enables psycholinguistic experts to select different sequential patterns as they appear in the translation process, compare the evolving changes of their statistics during the process and track their occurrences within a special simulator.