Figure 1.
The left side shows the components of the OQuaRE framework (Ontology Requirements, Quality Model and Quality Metrics) and the right side presents an example of traceability from ontology requeriments to metrics in OQuaRE.
Table 1.
OQuaRE Metrics.
Table 2.
Intervention of the study, ontology topics and modules.
Figure 2.
Significant effect of training in some topics.
22 subcharacteristics presented significant effect due to the GoodOD based training for some topics (PRO, IMM, CLO, CME, INF, SPA).
Figure 3.
No significant effect of training in any topic.
Seven OQuaRE subcharacteristics presented no significant effect of the training for any topic (PRO, IMM, CLO, CME, INF, SPA), and their mean values were similar for students and the gold standard.
Table 3.
P-values associated with the main effects in the two-way ANOVA with no effect of the interaction training by topic.
Figure 4.
Confidence intervals for Topic PRO: for trained and untrained students (left side); distances to gold standard (right side).
Table 4.
Significance levels of testing difference of means.
Figure 5.
Subcharacteristics classification for untrained and trained students, respectively.
Table 5.
Cluster centers.
Figure 6.
Factor loadings of topics in Principal Component Analysis.
Untrained case (left) and trained case (right). With the two more important components, more than 98% of total variance was explained for untrained students, and more than 92% of total variance for trained ones.
Table 6.
Sorting and classification of the 29 OQuaRE subcharacteristics for untrained (U) and trained (T) cases, respectively.