The quality of ontology learning results is usually compromised when a developer fails to recognize that a favored strategy for learning is actually uninformative and/ or a specific learning approach is applied to an unsuitable domain. Poor quality often results from a mismatch between learning strategies and domains. Given the importance of matching domains with methods to reuse [92], the selection of ontology learning approaches should take the characteristics of domains into consideration. Therefore, we classify domains from multiple perspectives and make recommendations of ontology learning approaches.