Time Series Analysis Using the Concept of Adaptable Threshold Similarity.pdf

Time Series Analysis Using the Concept of Adaptable Threshold Similarity ¨ Johannes Assfalg, Hans-Peter Kriegel, Peer Kroger, Peter Kunath, Alexey Pryakhin, Matthias Renz Institute for Informatics, University of Munich, Germany {assfalg,kriegel,kroegerp,kunath,pryakhin,renz}@dbs.ifi.lmu.de Abstract indication of heart desease T1 The issue of data mining in time series databases is of utmost importance for many practical applications and has attracted a lot of research in the past years. In this paper, we normal form focus on the recently proposed concept of threshold similar- ity which compares the time series based on the time frames T2 within which they exceed a user-defined amplitude thresh- old τ . We propose a novel approach for cluster analysis of time series based on adaptable threshold similarity. The most important issue in threshold similarity is the choice T3 of the threshold τ . Thus, the threshold τ is automatically adapted to the characteristics of a small training dataset using the concept of support vector machines. Thus, the op- timal τ is learned from a small training set in order to yield T1 T2 an accurate clustering of the en