Using path signatures to predict a diagnosis of Alzheimer’s disease

• DOI: 10.1371/journal.pone.0222212
• Publication Date: 2019-09-19T17:38:13Z
• Authors: P. J. Moore; T. J. Lyons; J. Gallacher
• Keywords: Signature (topology); Gold standard (test)

The path signature is a means of feature generation that can encode nonlinear interactions in the data as well usual linear features. It distinguish ordering time-sequenced changes: for example whether or not hippocampus shrinks fast, then slowly converse. provides interpretable features and its output fixed length vector irrespective number input points so it longitudinal varying with missing points. In this paper we demonstrate providing to set people Alzheimer's disease from matched healthy individuals. used are volume measurements whole brain, ventricles Disease Neuroimaging Initiative (ADNI). method shown be useful tool processing sequential which becoming increasingly available monitoring technologies applied.