What is STREAMER?

STREAMER is a modern framework that helps scientists to easily integrate and test machine learning algorithms into realistic streaming operational contexts.

Forget about dealing with distributed stream processing engines, STREAMER does it for you!

STREAMER is the first open-source academic framework that is scalable and cross-platform (Linux, Windows, Mac OS).

STREAMER is conceived to be used in two different ways (depending on your necessities):
a) Development use: (oriented to data scientists). You are interested on directly working on the code of the framework to add/develop several functionalities and test them.
b) Product use: (oriented to industrial use). You want to use the framework as a product (no need to get in contact with the code but execute STREAMER).

Experiment with your own use cases or algorithms, and use our API with available algorithms (clustering, classification...), preprocessing functions and evaluation metrics. Integrating your use case is quite simple and can be done in three steps:
(1) Define the data format to be used.
(2) Implement or call the algorithm to be tested.
(3) Configure the streaming context through the property files.

Take a look at this short presentation about STREAMER.

Download here the latest release (source code + environment + use case + algorithms)

Architecture

Demostrations: classical and distributed STREAMER

Funding Partners

Publications

  • Baudouin Naline, Sandra Garcia-Rodriguez Sandra, Karine Zeitouni. "STREAMER 3.0: Towards Online Monitoring and Distributed Learning". Proceedings of the 32nd ACM International Conference on Information and Knowledge Management. 2023. link
  • Sandra Garcia-Rodriguez, Mohammad Alshaer, and Cedric Gouy-Pailler. "STREAMER: a Powerful and Open-Source Framework for Continuous Learning in Data Streams." Proceedings of the 29th ACM International Conference on Information & Knowledge Management. 2020. link
  • Mohammad Alshaer, Sandra Garcia-Rodriguez and Cedric Gouy-Pailler. "Detecting Anomalies from Streaming Time Series Using Matrix Profile and Shapelets Learning." Proceedings of the 32th ACM International Conference on Tools with Artificial Intelligence. 2020. link
  • Other (peer-reviewed) scientific and divulgation publications: link