This project is an extension to the BeepBeep 3, event stream processing engine, called a palette, that provides functionalities to compute various statistical and basic clustering tasks on streams of events.
Please refer to the following research paper for detailed information and examples of what this extension can do.
M. Roudjane, D. Rebaïne, R. Khoury, S. Hallé. (2018). Real-Time Data Mining for Event Streams. Proc. EDOC 2018. DOI: 10.1109/EDOC.2018.00025. ResearchGate
Further documentation might be added to this Readme some time in the future (any help is welcome!).
To compile the palette, make sure you have the following:
- The Java Development Kit (JDK) to compile. The palette complies with Java version 6; it is probably safe to use any later version.
- Ant to automate the compilation and build process
The palette also requires the following Java libraries:
- The latest version of BeepBeep 3
- The latest version of Apache Commons Math 3
- Version 3.6.x of Weka (not the latest version, so that Java 1.6 is supported)
These three dependencies can be automatically downloaded and placed in the
dep folder of the project by typing:
ant download-deps
From the project's root folder, the sources can then be compiled by simply typing:
ant
This will produce a file called pattheminer.jar in the folder. This file
is not runnable and stand-alone. It is meant to be used in a Java project
alongside beepbeep-3.jar, commons-math3-x.x.x.jar and weka.jar (the JAR
files downloaded by the Ant build script).