ClaSP - Time Series Segmentation

This website contains supporting materials for our paper "ClaSP - Time Series Segmentation" published at CIKM'21. If you want to use ClaSP in your scientific publication or application, please use the updated and maintained claspy Python package.

Source Codes

ClaSP Code: A Python3 implementation of our code and all 98 datasets

FLOSS Code: A Python3 implementation of FLOSS

Autoplait Code: A C implementation of Autoplait

Ruptures Code: A  Python3 implementation of BinSeg-L2 / Window-L2

BOCD: A Python3 implementation of BOCD

The 98 Benchmark datasets

We wish to thank the data donors and archivists of the two benchmark sets for segmentation [1] and the UCR classification archive [2].

FLOSS Benchmark Datasets: 32 segmentation datasets (see [1]) used for the experiments

Semi-synthetic Benchmark Generator: A Jupyter Notebook to generate 66 semi-synthetic segmentation datasets from the UCR classification archive (see [2])

Dataset Description (Raw CSV, Raw Images): An overview of all 98 datasets

Results presented in the paper

Design Choices  (Raw CSV): The errors for all tested parameter combinations

Segmentation  (Raw CSV): The errors for ClaSP and its competitors

Runtime  (Raw CSV): The runtimes for ClaSP and its competitors

Window Size Selection  (Raw CSV): The errors for all tested window size selection strategies

Segmentation Visualizations (Raw Images): ClaSP segmentations for all 98 time series

Complex Time Series Classifiers (Raw CSV): The errors for ClaSP with different classifiers

References

[1] Supporting Website for "Domain Agnostic Online Semantic Segmentation": https://www.cs.ucr.edu/~eamonn/FLOSS/

[2] UEA & UCR Time Series classification archive: http://www.timeseriesclassification.com