12:00, Wed 26 Feb 1997, WeH 7220


		  Learning to Recognize Time Series:
	   Combining ARMA models with memory-based learning

			       Kan Deng
	
	In this talk, we propose a new method for classifying time
series data. Rather than compare features, as most prior approaches
do, we propose to compare the mechanisms underlying the generation of
a specific observation sequence, using the AutoRegression Moving
Average (ARMA) statistical model. For a given time series observation
sequence, we can estimate the parameters of the ARMA model, thereby
representing a potentially long time series by a limited dimensional
vector. In many applications, these parameter vectors will cluster
into different groups, based on the different mechanisms that generate
differing time series. We can then use classification algorithms to
predict the class of a new, uncategorized time series. For the
purposes of a highly autonomous system, our approach to this
classification uses memory-based learning and intensive
cross-validation for feature and kernel selection. In an example
application, we distinguish between driving data of a skilled, sober
driver vs. a drunk driver, by calculating the ARMA model for the
respective time series.