### Table 1. Subsumption tests for the reclassification of 16 concepts in TAMBIS.

2005

"... In PAGE 9: ...Table1 , which is divided into four columns. Each column reports the number of subsumption tests when using a certain form of approximation.... In PAGE 9: ...lassification (i.e., without approximation), the second column for C? i -approximation, the third column for C gt; i -approximation, and the fourth column for a combination of C? i - and C gt; i -approximation. Each column of Table1 is divided into a number of smaller rows and columns. The rows represent the level of the approximation used, where N denotes normal subsump- tion testing, i.... In PAGE 10: ...1 Analysis of C? i -/C gt; i -Approximation The approximation of concept classification in the TAMBIS ontology using the method of Cadoli and Schaerf reveals at least four points of interest. First, Table1 shows that using C? i -approximation can only lead to a reduction of the false subsumption tests and C gt; i -approximation can only lead to a reduction of the true subsumption tests. These results could be expected as they follow from Theorem 1 and are reflected by Algo- rithm 3, 4, and 5.... ..."

Cited by 6

### Table 1. Subsumption tests for the reclassification of 16 concepts in TAMBIS.

2005

"... In PAGE 9: ...Table1 , which is divided into four columns. Each column reports the number of subsumption tests when using a certain form of approximation.... In PAGE 9: ...lassification (i.e., without approximation), the second column for C? i -approximation, the third column for C gt; i -approximation, and the fourth column for a combination of C? i - and C gt; i -approximation. Each column of Table1 is divided into a number of smaller rows and columns. The rows represent the level of the approximation used, where N denotes normal subsump- tion testing, i.... In PAGE 10: ...1 Analysis of C? i -/C gt; i -Approximation The approximation of concept classification in the TAMBIS ontology using the method of Cadoli and Schaerf reveals at least four points of interest. First, Table1 shows that using C? i -approximation can only lead to a reduction of the false subsumption tests and C gt; i -approximation can only lead to a reduction of the true subsumption tests. These results could be expected as they follow from Theorem 1 and are reflected by Algo- rithm 3, 4, and 5.... ..."

Cited by 6

### Table 2. Number of subsumption tests for reclassification in five ontologies.

2005

"... In PAGE 12: ...For the first three Ontologies in Table2 , the DOLCE7, Galen8, and Monet on- tology9, C? i - or C gt; i -approximation has no effect. In these three Ontologies, C? i -/C gt; i - 7 An ontology for linguistic and cognitive engineering [18].... In PAGE 13: ...ions). Therefore no quantifiers were present in the ten randomly selected queries. As quantifiers are so rare, C? i -/C gt; i -approximation seems to be useless for those Ontologies. The next two Ontologies in Table2 , MadCow10 and Wine, are somewhat artificial because they are developed for demonstrating the expressive power of DLs. C? i -/C gt; i - approximation was applied to classification in both Ontologies, but this leads to almost no reduction of normal subsumption tests.... ..."

Cited by 6

### Table 2. Number of subsumption tests for reclassification in five ontologies.

2005

"... In PAGE 12: ...For the first three Ontologies in Table2 , the DOLCE7, Galen8, and Monet on- tology9, C? i - or C gt; i -approximation has no effect. In these three Ontologies, C? i -/C gt; i - 7 An ontology for linguistic and cognitive engineering [18].... In PAGE 13: ...ions). Therefore no quantifiers were present in the ten randomly selected queries. As quantifiers are so rare, C? i -/C gt; i -approximation seems to be useless for those Ontologies. The next two Ontologies in Table2 , MadCow10 and Wine, are somewhat artificial because they are developed for demonstrating the expressive power of DLs. C? i -/C gt; i - approximation was applied to classification in both Ontologies, but this leads to almost no reduction of normal subsumption tests.... ..."

Cited by 6

### Table 8: Subsumptions identified

2001

"... In PAGE 7: ... A total of 12 clusters (consisting of 10 articles each) were given to two judges each. Table8 indicates the num- ber of subsumptions found for each cluster. These numbers were computed by John Blitzer from Cornell University who is currently performing further analysis of the subsumption data.... ..."

Cited by 13

### Table 1. Levels of subsumption required to catch an exception Level of Subsumption Frequency

1999

"... In PAGE 10: ... All but two of the packages, namely the the Java JDK and servlet code, contain exception handlers that catch exceptions through subsumption. Table1 provides a di erent view of the data. This view illustrates that 32% of the di erent exception types present in try blocks remain uncaught in a target.... ..."

Cited by 1

### Table 2. Examples of subsumption conditions

1992

"... In PAGE 7: ... exception(W), the exception case in no compliance with general rules. Table2 shows examples of the SCs. 11 In the table, Ti and Tj are the POS sequences for the subsuming structure Ai and the subsumed structure Aj respectively.... ..."

Cited by 1

### Table 5: Incoherence and subsumption under an heterogeneous semantics

1997

"... In PAGE 16: ... The problem is now to detect incoherence and compute subsumption assuming an hetero- geneous semantics. The solution is given in Table5 where it can be seen that incoherence and subsumption relationships under an heterogeneous semantics can be turned into incoherence and subsumption relationships under the three given semantics. The relationships between virtual classes interpreted under the same semantics are imme- diate, thus we comment the ones between classes interpreted under di erent semantics: V l 2 vh V d 1 , V l 2 vl V d 1 and V l 2 vh V g 1 , V l 2 vl V g 1 ) : By adopting the heterogeneous semantics for the schema we also consider the lfp{instances of the schema .... ..."

Cited by 4

### Table 1. Levels of subsumption required to catch an exception

1999

"... In PAGE 10: ... All but two of the packages, namely the the Java JDK and servlet code, contain exception handlers that catch exceptions through subsumption. Table1 provides a di#0Berent view of the data. This view illustrates that 32#25 of the di#0Berent exception types presentintry blocks remain uncaught in a target.... ..."