New research paper on the use of BeepBeep for data mining

A paper describing an extension to the BeepBeep event stream processing engine has been accepted for publication at the 22nd International IEEE EDOC Conference, which will take place in Stockholm, Sweden, from October 16th to October 19th, 2018. The paper is co-authored by LIF Ph.D. student Massiva Roudjane, along with LIF faculty members Raphaël Khoury and Sylvain Hallé, and their UQAC colleague Djamal Rebaïne. EDOC 2018 is the twenty-second conference in a series that provides the key forum for researchers and practitioners in the field of enterprise computing.
 
The paper, titled Real-Time Data Mining for Event Streams, shows how trends of various kinds can be computed over logs in real time, using a generic framework called the “trend distance workflow”. Many common computations on event streams turn out to be special cases of this workflow, depending on how a handful of workflow parameters are defined. Experimental results presented in the paper indicate that deviations from a reference trend can be detected in real-time for streams producing up to thousands of events per second.
 
The extension is openly available on GitHub under a free software license at: https://github.com/liflab/PatTheMiner. BeepBeep itself is also available online and has an extensive documentation for beginners.
Posted in Research papers