Explaining why process instances are put in a particular cluster based on machine learning on features generated from control flow aspects.About -----
SECPI (Search for Explanations of Clusters of Process Instances) is technique that assists users with understanding a trace clustering solution by finding a minimal set of control-flow characteristics whose absence would prevent a process instance from remaining in its current cluster.
As such, the shortcoming of current trace clustering techniques regarding the provision of insight into the computation of a particular partitioning is addressed by learning concise individual rules that clearly explain why a certain instance is part of a cluster.
References
----------
SECPI has been described in the following contributions:
* De Koninck, P., De Weerdt, J., vanden Broucke, S. (2017). Explaining Clusterings of Process Instances. Data Mining and Knowledge Discovery 31 (3). pp. 774-808
* De Weerdt, J., vanden Broucke, S. (2014). SECPI: searching for explanations for clustered process instances. Proceeding of the 12th International Conference on Business Process Management, BPM 2014: Vol. accepted. International Conference (BPM 2014). Haifa (Israel), 7-11 September 2014.
Implementation
--------------
SECPI is implemented as a [ProM 6](http://www.promtools.org/) plugin. The following JAR file contains the plugin:
* [Version of 2016-02-16](downloads/svmexplainer-20160216.jar)
You will need to make sure that ProM can find the downloaded JAR in its classpath. To do so, you can create a folder “plugins” in the ProM installation directory, place the downloaded JAR file in this directory, and start ProM with the following command (Windows example):
java -classpath ./plugins/*;./lib/*;./* -Djava.util.Arrays.useLegacyMergeSort=true -Djava.library.path=./lib -ea -Xmx2g -XX:MaxPermSize=512m -XX:+UseCompressedOops org.processmining.contexts.uitopia.UI
In order for the SECPI-plugin to work correctly, a log file must be provided where each trace has a discrete `cluster:label` attribute assigned, denoting the cluster number the process instance belongs to.
When starting with a normal log file, you can rapidly create a valid clustered event log by first running the “ActiTraC” clustering algorithm. Next, using the created “ActiTraCClustering” object as input, apply the “Export Actitrac Solution to Clustered XES” plugin (included) to create a new, clustered event log. This clustered log file can then be used as input for the “Explain Clusters with SVM” plugins.
Alternatively, you can create a clustered log by applying the “Guide Tree Miner”-plugin. This results in a “ClusterLogOutput”-object. Next, you can use the “Export GuideTreeMiner Solution to Clustered XES” plugin (included) to create a new, clustered event log. Similarly, this clustered log file can then be used as input for the “Explain Clusters with SVM” plugins. If you wish to test SECPI on other process instance clustering techniques, you will need to create a clustered event log (with the `cluster:label` attribute) via other means.
A plugin to perform a grid optimization to retrieve the optimal c and eps parameters for the SVM is also included.
Source code is available on [GitHub](https://github.com/Macuyiko/processmining-secpi).
Contact
-------
Contact the authors at:
* [Jochen De Weerdt](mailto:jochen.deweerdt@kuleuven.be) (corresponding author)Department of Decision Sciences and Information Management, KU Leuven
Naamsestraat 69, B-3000 Leuven, Belgium * [Pieter De Koninck](mailto:pieter.dekoninck@kuleuven.be) (corresponding author)
Department of Decision Sciences and Information Management, KU Leuven
Naamsestraat 69, B-3000 Leuven, Belgium * [Seppe vanden Broucke](mailto:seppe.vandenbroucke@kuleuven.be) (corresponding author)
Department of Decision Sciences and Information Management, KU Leuven
Naamsestraat 69, B-3000 Leuven, Belgium Screenshots ----------- [
](#i00)
[
](#i01)
[
](#i02)
[
](#i03)
[
](#i04)
[
](#i05)
[
](#i06)
[
](#i07)
[
](#i08)
[
](#i09)