Performance Metrics in Ontologies
Description

Semantic performance analysis shows performance metrics for concepts in ontologies. The current version of this plug-in supports two metrics: processing times (cf. Figure 1) and throughput times (cf. Figure 2). This plug-in has the following functionalities:

  1. Visualization of throughput times. This metric shows how much time process instances of a certain concept take.
  2. Visualization of processing times. This metric shows the amount of time originators of certain concepts spent on work or how much time tasks of certain concepts took to be processed.
  3. Visualization of performance metrics per ontologies. Each ontology in the log to which processing/through times could be calculated has a tab in the main panel Ontology. For instance, Figure 1 contains two tabs: NexcomActivities and OrganisationalOntology. Note that the tab NexcomActivities is selected in this figure and a projected view of this ontology is shown in the middle panel. The colors of the concepts allow the end user to quickly spot potential bottlenecks in the system. Dark green means fast and red means slow (with gradations in between).
  4. Filtering of which direct instances to take into account during the calculation of the performance metrics. This functionality if provided at the left panel Filtering. If a concept is selected in this panel, all their direct instances are taken into account. Otherwise, these instances are not used during the calculation of the performance metrics. For instance, in Figure 1 the concepts OrganisationalOntology#Director and OrganisationalOntology#CEO are deselected. Since instances of these concepts execute tasks of the concept NexcomAcitivites#RequestSupplierOfferAdvice, you can see (cf. Ontology panel) that no processing time could be calculated for these concepts.
  5. Visualization of the actual metric numbers in a tabular way. This functionality is provided by the top-right panel Numbers. The colors in the rows of the tables are the same ones used for the ontology in the Ontology panel. The table has four columns with the following meanings: (a) Name: name of the concepts in the visualized ontology; (b) Total: performance metric as an absolute value; © Frequency: number of included measurements; and (d) Average: Total divided by Frequency.
  6. Exportation of the actual metric numbers. The table explained in the previous step can be exported to the CVS format used by MS Excel.
  7. Selection of different units for the visualization of the performance metrics. This functionality is provided at the bottom-right panel where the option Number format is available.
  8. Selection of which column of the performance table to use when coloring the ontology and the table. This functionality is provided at the bottom-right panel where the option Colors based on is available.
Figure 1. Screenshot of the plug-in Performance Metrics in Ontologies. Option Show processing times.
Figure 2. Screenshot of the plug-in Performance Metrics in Ontologies. Option Show throughput times.
How to use
  1. Open a SA-MXML log file by clicking on File→Open Supported File….
  2. Start the Performance Metrics in Ontologies plug-in via clicking Analysis→ LogName→ Performance Metrics in Ontologies. You should get a screen like the one in Figure 1.
  3. If you like, you can export the table with the performance metrics to CSV format by clicking the menu option Exports→Table with values→Standard CSV(Excel).