This repository contains code to sample time series efficiently from a large-scale collection, such that the sampled series can cover the series shapes in the collection.

Install as a Python module:

python setup.py build_py_mod 
pip install .

If you need a shared library, run:

python setup.py build_lib 

An example usage Jupyter notebook is provided at example/seasam_example.ipynb

@inproceedings{conf/kdd/WangP21,
  author       = {Qitong Wang and
                  Themis Palpanas},
  title        = {Deep Learning Embeddings for Data Series Similarity Search},
  booktitle    = {{KDD}},
  pages        = {1708--1716},
  year         = {2021},
}

If you find SEAsam sampling useful, please make sure to cite also the original SEAnet paper (where SEAsam is proposed): [bibtex]