China
Africa
Explore chapters and articles related to this topic
Knowledge Representation and Storage
Published in Sarika Jain, Understanding Semantics-Based Decision Support, 2021
Various tools, such as OntoMetrics, OntoClean, and OntoQA, are available for evaluating ontologies quantitatively. The EO and TO were evaluated with the OntoMetrics tool to calculate their statistics. Table 4.3 shows the values of the metrics for both ontologies, collectively showing how much knowledge is encoded in each ontology.Base metric: The base metric includes simple parameters such as class count (CC), axiom count (AC), object property count (OPC), data property count (DPC), and individual count (IC), among many more. The tools in Protégé and OntoMetrics were employed to measure these various parameters.Schema metric: The design of the ontology is evaluated by the schema metric. This metric addresses attribute richness (AR), inheritance richness (IR), relationship richness (RR), the axiom/class ratio (A/C R), the inverse relation ratio (IRR), and the class/relation ratio (C/R R). AR is the average number of attributes per class. IR shows how well knowledge is organized into different groups and subgroups in the ontology. RR reflects the heterogeneity of the relationships in the ontology; it is calculated by dividing the number of non-inheritance relationships by the total number of relationships defined in the schema. A/C R represents the average amount of axioms per class. IRR is the ratio of inverse relations to total relations. C/R R is the ratio of classes to relations in the ontology.Instance metric: The instance metric evaluates how the specified knowledge is utilized, providing clues about the effectiveness of the ontology design. Two parameters are calculated: average population (AP) and class richness (CR). AP specifies the number of individuals as compared to the number of classes. CR indicates how instances are arranged across the classes, i.e., comparing between different classes.Graph metric: The graph metric evaluates the structure of the ontology according to parameters like absolute root cardinality (ARC), absolute leaf cardinality (ALC), sibling cardinality (SC), depth (D), and breadth (B).Class metric: The class metric examines the classes and relationships of the ontology by measuring class importance (CI), class readability (CR), and class instance count (CIC) for every class. For the EO, these metrics were measured for the class “Action”; for the TO, “Armaments.”
Development and evaluation of knowledge treasure for emergency situation awareness
Published in International Journal of Computers and Applications, 2021
This technique offers a quantitative perspective to evaluate ontologies. This approach has been adopted by several techniques. Reviews of these techniques have been done in this section. OntoClean: It follows a feature-based approach [37] to evaluate and validate the ontology. This technique assigned the four features (Rigidity, Unity, Identity, and Dependence) to each of the classes in ontology to identify the problematic areas. Classes can be moved up or down in the taxonomy and new classes can be added or removed from the ontology to correct the problems identified through the detection of the offense of a set of rules based on the four features.AKTiveRank: Alani [38] presents a method known by AKTiveRank which identify the related ontologies based on the terms entered by the user. It evaluates the schema to select the most suitable ontology. The measures which are developed by AKTiveRank are: class match, density, semantic similarity, and betweenness.oQual: Gangemi et al. [39] developed this technique which evaluates the ontology on three dimensions: Structural in which 32 features are used to analyze the syntax and semantics of the ontology; Functional in which five qualitative measures are used to analyze the relationship between ontology and its meaning; and third is usability profiling in which main focus is on the context of the ontology.Ontometric: It is a hierarchical framework which consists of 160 features to evaluate the suitability and quality of ontologies according to the user's system requirements [40]. These features are spread across five dimensions such as the content of the ontology, building tools, development methodology, language, and usage costs. It categorized the metric into Base metric, Schema metric, instance metric, and class metric.ONTOQA: It is metric-based ontology evaluation tool [36]. Metric is divided into two groups: schema metrics and knowledge base metrics.