Project 2-SPARQL library

[elenamiliv]

This assignment outlines a series of SPARQL queries designed for evaluating and analyzing ontology structure and quality. Each query addresses specific aspects of ontology evaluation, identifying potential issues or improvements necessary to enhance clarity and consistency.

1. Detect Properties Without Domain or Range (level 8 - 0 points)
2. Identifying Disjoint Classes with Common Instances (level 7 - 2 points)
3. Detecting Unused Object Properties (level 7 - 2 points)
4. Finding Classes with Conflicting Disjointness and Subclass Relationships (level 6 - 3 points)
5. Identifying Classes with Different Labeling Standards (level 6 - 3 points)
6. Identifying Potentially Overspecified Class Definitions (level 5 - 5 points)
7. Detecting Equivalent Properties with Different Ranges (level 5 - 5 points)
8. Analyzing Class Hierarchies for Inconsistent Type Definitions (level 5 - 5 points)
9. Detecting Potentially Misplaced Inverse Object Properties (level 5 - 5 points)
10. Identifying Classes with No Properties (Level 4 - 10 points)
11. Identifying Classes Without Labels (Level 4 - 10 points)
12. Query to Detect Missing Annotations (Labels or Comments) Level 4 - 10 points
13. Detecting Properties Shared Across Multiple CCO Classes Level 3 - 20 points
14. Detecting Inconsistent Property Ranges Across Subclass Hierarchies Level 3 - 20 points

[johnbeve]

Where?

[johnbeve]

Suppose you have a class A that is asserted in the OWL file to be owl:disjointWith itself. Suppose it has an instance a. This query would return: A, A, a. This is problematic and should be detected, but not for the motivation of this query.
This suggests to me you should assert ?class1 =/ ?class2 and have a separate query that identifies as errors classes asserted to be disjoint with themselves.

[johnbeve]

First, the classes could be the same as discussed above.
Second, I'm not sure how this query does what you describe, since it's unclear what standards are being leveraged in evaluation.

[johnbeve]

You need to leverage the specific annotation property used in CCO for definitions to narrow the focus of your query. Everything with a ? preceding it is a variable and so by default ranges over everything that might match its position in the triple.

[johnbeve]

I need to know much more about what you take this query to be doing...

[johnbeve]

And you think this is worth 10 points?

[johnbeve]

Should all CCO elements have comments?

[johnbeve]

Compare this against the following and report back the results:

PREFIX rdf: http://www.w3.org/1999/02/22-rdf-syntax-ns#
PREFIX rdfs: http://www.w3.org/2000/01/rdf-schema#

SELECT DISTINCT ?property ?class1 ?class2
WHERE {
?property a rdf:Property .

?instance1 a ?class1 .
?instance1 ?property ?value .

?instance2 a ?class2 .
?instance2 ?property ?value .

FILTER(?class1 != ?class2)
}

[johnbeve]

First, certainly not worth 20 points.
Second, compare and report back the results:

PREFIX rdf: http://www.w3.org/1999/02/22-rdf-syntax-ns#
PREFIX rdfs: http://www.w3.org/2000/01/rdf-schema#
PREFIX owl: http://www.w3.org/2002/07/owl#

SELECT DISTINCT ?property ?class1 ?range1 ?class2 ?range2
WHERE {
?property a rdf:Property .

?class1 ?property ?value1 .
?value1 a ?range1 .

?class2 rdfs:subClassOf ?class1 .
?class2 ?property ?value2 .
?value2 a ?range2 .

FILTER(?range1 != ?range2)
}