The structure of domain entity attribute semantic knowledge base for natural language processing

The abstraction of semantic knowledge in terms of entities, attributes, and action relationships can completely define two semantic forms of knowledge: picture and situation. The knowledge base proposed in this paper contained the entries that correspond to the components of the real world and the organizational network in which those entries are associated with the situation. The knowledge base contains situation verb base, entity base, attribute base and attribute value base. Those terms are organized into a network according to the external feature of the entity and the relationship between entities and attributes. The knowledge base also defines two sets of knowledge networks. One is the entity attribute knowledge frame which is integrated by the entity as a unit. It defines the frame of an entity corresponding to the real object. It embeds the slot into the entity, and points out the possible values of each attribute slot. The other is the collocation frame which is integrated by the verb as the core unit. The knowledge resource is stored in XML format which could hierarchically represent the structure of knowledge framework. It could support the addition and deletion of knowledge resources making the continuous construction and expansion of the knowledge base possible. The abstracts of knowledge in this paper reproduced the complete process of language comprehension, which can be well applied to language analysis.