Finding approximate frequent patterns in streaming medical data

Time series data is ubiquitous and plays an important role in virtually every domain. For example, in medicine, the advancement of computer technology has enabled more sophisticated patients monitoring, either on-site or remotely. Such monitoring produces massive amount of time series data, which contain valuable information for pattern learning and knowledge discovery. In this paper, we explore the problem of identifying frequently occurring patterns, or motifs, in streaming medical data. The problem of frequent patterns mining has many potential applications, including compression, summarization, and event prediction. We propose a novel approach based on grammar induction that allows the discovery of approximate, variable-length motifs in streaming data. The preliminary results show that the grammar-based approach is able to find some important motifs in some medical data, and suggest that using grammar-based algorithms for time series pattern discovery might be worth exploring.