A0185
Title: Linear law-based feature space transformation
Authors: Marcell Tamas Kurbucz - Wigner Research Centre for Physics | Corvinus University of Budapest (Hungary) [presenting]
Antal Jakovac - Wigner Research Centre for Physics (Hungary)
Peter Posfay - Wigner Research Centre for Physics (Hungary)
Abstract: The aim is to facilitate uni- and multivariate time series classification tasks using linear law-based feature space transformation (LLT). This new algorithm first splits the instances into training and test sets. Then, it identifies the governing patterns (laws) of each input sequence in the training set by applying time-delay embedding and spectral decomposition. Finally, it uses the laws of the training set to transform the feature space of the test set. These calculation steps have a low computational cost and the potential to form a learning algorithm. The application of LLT is illustrated using datasets from the fields of human activity recognition, price movement prediction, and electrocardiogram signal classification.
