General architecture

In this chapter, we discuss the proposed architecture for classification in temporal domains. We discuss the training architecture, which takes the examples given and tries to extract rules for describing time series, as well as the testing, or recognition architecture, which can take the outputs from the training architecture and apply them efficiently to unseen streams.

• Overview
• Training architecture
  • Raw training data
  • Parametrised event primitives
    • Example PEP: A ``straight-line'' approximation for continuous-valued channels
    • Example PEP: ``Delta'' detection for discrete-valued attributes
  • Event extraction
  • Global features
    • Example global attributes
  • Parameter clustering
  • Event attribution
  • Recombination
  • Feature selection
  • Conventional attribute-value learning
  • Classifier
• Testing architecture
  • Raw test instance
  • Selected event search
  • Selected global feature calculation
  • Recombination
  • Classifier
• Goal evaluation
  • Generality
  • Classification accuracy
  • Meaningful descriptions
  • Fast learning curve
  • Time