|
Abstract : |
Finding patterns in temporally structured data is an important and di cult problem. Examples of temporally structured data include time series of economic indicators, distributed network status reports, and continuous streams such as ight recorder data. We have developed a family of algorithms for nding structure in multivariate, discrete-valued time series data (Oates & Cohen 1996b ??? Oates, Schmill, & Cohen 1996 ??? Oates et al. 1995). In this paper, we introduce a new member of that family for handling event-based data, and o er an empirical characterization of a time series based algorithm. 1, |
