Algebraic time series forecasting and segmentation models
Kristina Lukoševičiūtė, Rita Palivonaitė · Lietuvos matematikos rinkinys · 2015
An algebraic segmentation method based on algebraic predictor with internal smoothing is proposed. The concept of the rank of the sequence is proposed for the detection of a base fragment of the sequence. Numerical experiments with a real-world financial time series illustrate the segmentation method.