Core Assessment Vocabulary

Abstract

Introduction back to ToC

The Core Assessment Vocabulary represents, expresses, and defines what an “Assessment” of “Assets” is and how to perform the assessment based on “Criteria”. It is a domain-agnostic vocabulary, meaning that it can be used to assess any asset. The CAV addresses semantic interoperability by reusing existing generic ontologies and vocabularies. This way, the semantics of common concepts and properties are agreed without having to re-discuss them. When concepts or properties have not been identified nor defined for the purposes pursued, they have to be proposed either as extensions or from scratch.

The methodological approach followed for the development of the CAV reuses the following ontologies and vocabularies:

• Data Catalogue Vocabulary (DCAT)
• Friend of a Friend (FOAF)
• Asset Description Metadata Schema (ADMS)
• The Organization Ontology
• Core Criterion and Core Evidence Vocabulary (CCCEV)
• DCMI Metadata Terms (DCTerms)
• Schema

Core Assessment Vocabulary: Overview back to ToC

Classes

• Assessment
• Asset
• Dataset
• Evaluation context
• Scenario
• Score
• Statement

Object Properties

• Assesses
• Assigned to
• Based on
• Considers
• Contact point
• Based on
• Contextualises
• Defined by
• Defines
• Defines sub
• distribution
• Documented in
• Formulated by
• Has input
• Has report
• Has summary
• Includes
• Performed by
• Previous version
• Provided by
• Provides score
• Rated in
• Refers to
• Relates to
• Requested by
• Results in
• Triggered from
• Triggers

Data Properties

• Description
• Instruction description
• Interpretation
• Issued
• Judgement
• Purpose
• Release date
• Status
• Title
• Type
• Value
• Value description
• Version

Core Assessment Vocabulary: Description back to ToC

A CAV Assessment is a specialisation of an Asset, which is itself also a specialisation of the dcat:Dataset and by extension of dcat:Resource.

RDFS(RDF Schema) provides a basic vocabulary for defining classes and properties of a resource. Within this framework, an rdfs:Resource represents the most general concept, capable of describing anything.

CAV, however, needs to clearly delimit which resources are subject to assessment, and it does so by relying on widely accepted standards. DCAT builds on the abstract notion of a resource and makes it operational for data catalogues. Since the subject of an assessment is intended to be a dataset or a service, a cav:Assessment assesses a dcat:Resource.

In parallel, ADMS adopts a more general perspective, allowing an rdfs:Resource to describe its own content; in this case, the resource is implicitly represented as an adms:Asset. Because an adms:Asset can be catalogued —and is therefore considered a dcat:Dataset, and because a cav:Assessment is itself meant to be catalogued, it follows logically that any assessment of a resource becomes an asset. This ensures full alignment with the ADMS specification.

This alignment is important because CAV assessments are published in catalogues like the Interoperable Europe Portal. Using adms:Asset helps CAV avoid repeating existing concepts, keeps the vocabulary consistent, and makes it easier to connect with other frameworks. As a result, CAV assets will be preserved on the basis of documenting and making them compatible with CCCEV.

As any ADMS Asset, the Assessment can be identified and described, and has individuals’ distributions, publishers, etc. Additionally to these properties the CAV class Assessment needs to specify who are the requestors and evaluators of the Assessments. These can be anything represented by a FOAF Agent, such as a natural person, a group or an organisation. Notice that the objects assessed are also Resources meaning that the CAV may be used to assess anything that is considered a valuable resource . Example of such resources could be products, services, or, in the case of CAMSS, standards and specifications.

An Assessment results in Statements capturing the produced knowledge and providing value judgments. These can refer to the Assessment as a whole or a specific section, even being as detailed as to refer to individual evaluated criteria.

A Criterion is typically derived from a Reference Framework, which is to be understood as a series of “agreed and descriptive reference requirements” coming from one or more sources (e.g., legislation, specifications and standards, ICT policy-related works like the EIF within the EIS, etc.). Throughout the Assessment each Criterion is assigned a Score (in principle by humans, but potentially also by systems) as the value output that is considered when formulating the resulting Statement(s). The Score can take into consideration any resource input (e.g., when the score is calculated based on different input parameters, algorithms, and formulae).

Any Assessment is performed in the context of a Scenario. The Scenario defines the purpose of the Assessment and the set of Criteria to be scored by one or more Agents. Scenarios can be defined with a flexible structure including nested sections (represented as sub-Scenarios) that serve to provide additional context, group thematically Criteria and be referred to by the assessment’s resulting Statement(s). Criteria can themselves be simple or complex and originate from various reference sources. The overall context for the evaluation of the Criteria is provided by the Scenario, however in case certain Criteria require additional contextualisation or evaluation instructions these can exceptionally be provided by means of Evaluation Context which is related to a resource . Moreover, an Assessment might trigger another related Assessment of different content which has its own Scenario and Criteria. Finally, an Assessment has a title and a description to facilitate the identification of the Asset. Note that it is also possible to model work in progress, expressed by having the Assessment defining optional links to Scores, Statements, reports and summaries. The versioning of an Assessment is represented by the version datatype property of DCAT to keep track of the current version of the Assessment and its relationship with a previous one; setting up the issued date (assessment performance date) and the release date (assessment publication date) of the Assessment is also recommended, where the CAV reuses the DCTerms issued (date) property for both.