Abstract

The Kolmogorov-Zurbenko Adaptive, kza package provides algorithms to deal with abrupt changes or breaks in the presence of heavy background noise. In a practical way, one-dimensional and high-dimensional simulated samples are generated to demonstrate signal recoveries and their accuracy evaluation by mean of squared error, mean difference and specificity index. Simulation investigation showed that smoothing window size need consider whenever applying kza package, and that kza could tolerate background noise about 10-folds heavier in higher-dimensional data compared to 1-dimensional data.

This work is licensed under a Creative Commons Attribution 4.0 License.