Scalar attributes and time series combining patterns for multivariate time series classification

Abstract

In this paper, we propose new patterns which combine attribute and time series data for classifications to soften or solve the overlapping classes problem. The solution was created for medical human gait analysis, where both types of input data are available.