With domain ontology, a meaningful index of document indexing, such as the domain events structure in this paper, can be defined. Since the construction of domain ontology is costly, an automatic domain ontology acquisition is preferred. We find that named entity can be a clue of domain event acquisition from the training corpus. Therefore, we add the named entity recognition module into the automatic domain ontology acquisition process. Thus, a subject-verb-object-modifier (SVOM) indexing can be constructed. Our experimental results demonstrate that the automatic event extraction and ontology acquisition can be good resources for text categorization and further information processing.