OWLSummarizer

1. Description

OWLSummarizer is a tool to produce automatic summaries of OWL ontologies. It is implemented in Java and uses the OWL API to manipulate ontologies. Basically, the tool can be invoked at command line and accepts a set of input parameters. The parameter values can be adjusted at the summary.properties file. An optional log file is generated after each successful execution. The log file contains useful information for user analysis. OWLSummarizer is under development at the Center for Informatics of the Federal University of Pernambuco, Recife, PE, Brazil.

2. Using OWLSummarizer

The main steps to use OWLSummarizer are:

a) Uncompress OWLSummarizer.zip to an empty folder;

b) Configure summary.properties;

c) Execute summarize.bat at the DOS command line interface.

Requirement: a JRE (virtual machine) must be defined at the path environment variable of Windows.

Example:

C:\OWLSummarizer\summarize.bat example.owl exampleSummary.owl

If the execution is successful, two files are created: an OWL file containing the ontology summary (exampleSummary.owl) and a log file (example.log). The log file contains useful information produced internally by the tool during the execution, e.g. elapsed time and identified subontologies. It can be used by a user for subsequent analysis. Table 1 describes each input parameter and shows their respective default values.

The complete syntax is:

summarize <owl_filename> <owl_summarized_filename> <frenquency_filename> <centrality_weight> <frequency_weight> <alpha_measure> <relevant_concepts> <summary_size_reference> <summary_size_delta>

where:

<owl_filename> = OWL file containing the ontology to be summarized. The file must be at the same folder of summarize.bat;

<owl_summarized_filename> = OWL file representing the ontology summary;

<frequency_filename> = XML file containing the frequency of the concepts. The file must be at the same folder of summarize.bat;

<centrality_weight> = Centrality weight. Example: 0.5. Default: 1.0;

<frequency_weight> = Frequency weight. Example: 0.5. Default: 0.0;

<alpha_measure> = Determines the importance of recall and precision (F-measure). Example: 0.5. Default: 0.5;

<relevant_concepts> = Number of relevant concepts in an ontology. Example: 12. Default: 6;

<summary_size_reference> = Suggested size for an ontology summary. Example: 12. Default: 6;

<summary_size_delta> = Variation of the size of the suggested ontology summary. Example: 2. Default: 2.

Useful hints:

It is not necessary to inform all parameters at command prompt. Instead, the parameters can be initialized at the summary.properties file. The file is located at the OWLSummarizer folder.

When the numeric parameters are not informed at command prompt, the values defined at the summary.properties files are considered.

When the frequency file is not informed, the frequency measure is ignored during the summarization process. In other words, only the centrality measure is considered.

3. Ontology Examples

In this section, we present the ontologies which were used during the tests: conference.owl, networkA.owl, office-env2.owl, and univ-cs.owl. All of them are third-party ontologies which are available in the Web.

Table 1. Ontology information.

Ontology Name	Knowledge Domain	Description	# Classes	# Properties
conference.owl	Conference	An ontology draft for events and, specifically, conferences	18	18
networkA.owl	Network	Describes the nodes in a local area network	27	6
office-env2.owl	Office	Models an office environment	35	8
univ-cs.owl	University	Describes a computer science academic department	53	25

Figure 1. Conference ontology.

Figure 2. Network Ontology.

Figure 3. Office Environment Ontology.

Figure 4. University Ontology.

4. Downloads

Tool: OWLSummarizer.zip

Ontologies: ontologies.rar

5. Contact Info

Victor Alencar - vba [at] cin [dot] ufpe [dot] br

Carlos Pires - cesp [at] cin [dot] ufpe [dot] br