
178 ◾ Temporal Data Mining
series. e authors proved the eciency of their method on both real and
synthetic data sets.
5.4.5 Complex Temporal Pattern Identification
In [Hon02], the authors discuss a method for the extraction of temporal
patterns in a signal that is also dominated by noise and nonlinear temporal
warping. e main idea of the algorithm is to use a hidden Markov model
(HMM). e proposed approach detects patterns, even if the number and
length of patterns are unknown and the sequence contains irregular non-
pattern signals. A specic type of HMM is used. It is a threshold HMM.
e thresholds are placed on the population of the ...