iKnow Terminology

Term	Explanation
Attribute	This term usually refers to a semantic attribute (Negation, Sentiment, Time etc.), but within the language model it can refer to a typeAttribute label. TypeAttribute labels can be used for all sorts of information that does not affect the role (Concept - Relation - PathRelevant - NonRelevant) of a lexrep
Concept	a type of entity, typically a noun phrase
Dominance	a relevance metric generated by factoring in the entity frequency relative to the length of each text, the frequency of other entities having words in common, and other factors. Plain TFIDF scores would otherwise not take into account words shared between entities. Refer to InterSystems IRIS documentation for more detail on Dominance.
Proximity	a metric expressing how related two concepts are, based on their relative distance in a given text or corpus. For Japanese, Proximity is based on the entity order within the Entity Vector, not the original sentence entity order. Refer to InterSystems IRIS documentation for more detail on Proximity.
Entity	a word or sequence of words that iKnow outputs as a unit
Entity Vector	an alternative for Paths, used in the Japanese language model. The order of the entities in an Entity Vector reflects their importance in the sentence, which is not necessarily the order in which they occur in the input.
Lexrep	short for 'lexical representation', i.e. a word or sequence of words that is treated as a unit in the linguistic processing (language model)
NonRelevant	a type of entity, representing a lexrep that does not express an essential part of the meaning of a sentence, e.g. an article
Path	a sequence of Concepts, Relations and PathRelevant entities, used as a level to illustrate the context of an entity and the scope of attributes
PathRelevant	a type of entity, representing a lexrep that doesn't have a fixed referent without a context, e.g. a personal pronoun, vague time expression or grammatical place holder
Relation	a type of entity, typically a preposition, conjunction or verb