Generalized Chronicles for Temporal Sequence Classification
Abstract
Discriminant chronicle mining (DCM) tackles temporal sequence classification by combining machine learning and chronicle mining algorithms. A chronicle is a set of events related by temporal boundaries on the delay between event occurrences. Such temporal constraints are poorly expressive and discriminant chronicles may lack of accuracy.
This article generalizes discriminant chronicle mining by modeling complex temporal constraints. We present the generalized model and we instantiate different generalized chronicle models. The accuracy of these models are compared with each other on simulated and real datasets.
Origin : Files produced by the author(s)