Table 2.
Characteristics of nine existing ontology learning systems
| Input | Language | Ontological Elements Learned | Degree of Automation | Resource | Ontology Enrichment or De Novo generation | Learning Methods | |
|---|---|---|---|---|---|---|---|
| ASIUM | Free text documents. | French | Concepts, taxonomic relations | Semi-automated | N/A | Deno Vo | Conceptual and hierarchical clustering |
| DODDLE II | Dictionary, domain specific text documents | English | Concepts, taxonomic relations, non-taxonomic relations | Semi-automated | WordNet | Enrichment | Matching and trimming against WordNet for taxonomic relations, statistical co-occurrence information |
| HASTI | Free text documents | Persian | Concepts, taxonomic relations, non-taxonomic relations, axioms | Two modes: semi-automated and fully-automated | N/A | Deno Vo | Combination of logical. linguistic, template, and heuristic |
| KnowItAll | Web pages | English | Concepts, | Automatic | Domain ontology | Enrichment | Combination of linguistic and statistic methods |
| MEDSYNDIKATE | Medical domain documents | German | Concepts, taxonomic relations | Semi-automated | Own general and medical lexicons, Fully lexicalized dependency grammar | Enrichment | Input text is mapped to corresponding text knowledge bases (TKB) which represent the text content, Generates concept hypothesis and ranks hypothesis based on quality |
| OntoLearn | Free text documents | English | Concepts, taxonomic relations | Semi-automated | WordNet, SemCor | Enrichment | Machine learning, Statistical approach |
| STRING-IE | Free text documents from PubMed | English | Non-taxonomic relations | Automated | SWISS-PROT, Saccharomyces Genome Database | Enrichment | Linguistic and rule based approach |
| Text-To-Onto Text2Onto |
Dictionaries, Databases, Semi-structured text, Free text documents | German | Concepts, taxonomic relations, non-taxonomic relations, | Semi-automated | Domain ontology (Tourism) | Enrichment | Combination of association rules, formal concept analysis and clustering |
| TIMS | Free text documents | English | Concepts, taxonomic relations | Automated | N/A | Enrichment | Automatic term recognition using both linguistic and statistical approach and automatic clustering using average mutual information |
| WEB → KB | Web pages | English | Concepts, taxonomic relations | Automated | Domain ontology | Enrichment | Statistical and Logical |