Skip to content

Heuristics Miner

Dependency

To use these algorithms in your Java Maven project it is necessary to include, in the pom.xml file, the dependency: 

<dependency>
    <groupId>com.github.beamline</groupId>
    <artifactId>discovery-heuristics</artifactId>
    <version>master-SNAPSHOT</version>
</dependency>
See the introduction page for further instructions.

Usage

This miner contains two version of the streaming Heuristics miner. One is based on the Lossy Counting algorithm, the other is based on the Lossy Counting with Budget. These can be accessed with the following parameters:

1
2
3
4
5
6
HeuristicsMinerLossyCounting miner = new HeuristicsMinerLossyCounting(
    0.0001, // the maximal approximation error
    0.8, // the minimum dependency threshold
    10, // the positive observation threshold
    0.1 // the and threshold
);

1
2
3
4
5
6
HeuristicsMinerBudgetLossyCounting miner = new HeuristicsMinerBudgetLossyCounting(
    100000, // the total budget available
    0.8, // the minimum dependency threshold
    10, // the positive observation threshold
    0.1 // the and threshold
);

After the miner is configured, both can be used to produce a CNet which can be either exported into a .cnet file or visualized (currently the visualization does not support the bindings):

 7
 8
 9
10
11
12
13
14
15
16
17
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
env
    .addSource(source)
    .keyBy(BEvent::getProcessName)
    .flatMap(miner.setModelRefreshRate(100))
    .addSink(new SinkFunction<StreamingCNet>(){
        public void invoke(StreamingCNet value, Context context) throws Exception {
            new CNetSimplifiedModelView(value.getCnet()).exportToSvg(new File("output.svg"));
        };
    });
env.execute();

Scientific literature

The techniques implemented in this package are described in: