Amit Mathew

1/27/04

CSCI 8351

WordNet

 

According to its web page, WordNet is a lexical reference system.  WordNet provides an implementation of a system that associates English words in several different ways.  Words are associated by synonyms, hypernyms, hyponyms, coordinate terms, meronyms, derivationally related forms, domain, and familiarity.  Hypernyms define a "is-a" relationship.  For example, a vehicle is a hypernym of car means that a car is a type of vehicle.  A hyponym is the "superset" relationship.  For example, a car is a hyponym of vehicle, because a car is a type of vehicle.  Coordinate terms are words with the same hypernym.  Meronyms is the name for a constituent part, the substance of, or a member of something.  For example, a student is a meronym of school because a student is a member of a school.  Derivationally related forms are nouns and verbs that are related morphologically.  In other words, derivationally related forms provide links between the otherwise separate classes of nouns and verbs.  An example of this is "a cover" and "to cover."  Domain is used to classify objects by category, region, or usage and provides a method of differentiating words whose meaning changes depending on the area, like the word "ontology," which differs in the philosophy and information theory domains.  Familiarity refers to the polysemy count, or the number of senses the word has in WordNet.  Each sense refers to a different definition of each word.

 

I inputted several nouns in to the online WordNet database.  For this write-up, I use the word "net."

 

For the noun "net," I got six senses:

 

·        Internet, Net, cyberspace -- (a computer network consisting of a worldwide network of computer networks that use the TCP/IP network protocols to facilitate data transmission and exchange)

·        net -- (a trap made of netting to catch fish or birds or insects)

·        net income, net, net profit, lucre, profit, profits, earnings -- (the excess of revenues over outlays in a given period of time (including depreciation and other non-cash expenses))

·        net -- (a goal lined with netting (as in soccer or hockey))

·        net -- (game equipment consisting of a strip of netting dividing the playing area in tennis or badminton)

·        net, network, mesh, meshing, meshwork -- (an open fabric of string or rope or wire woven together at regular intervals)

 

I also retrieved several synonyms:

 

·        computer network -- ((computer science) a network of computers)

·        trap --(a device in which something (usually an animal) can be caught and penned)

·         income -- (the financial gain (earned or unearned) accruing over a given period of time)

·        goal -- (game equipment consisting of the place toward which players of a game try to advance a ball or puck in order to score points)

·        game equipment -- (equipment or apparatus used in playing a game)

·        fabric, cloth, material, textile -- (artifact made by weaving or felting or knitting or crocheting natural or synthetic fibers; "the fabric in the curtains was light and semitraqnsparent"; "woven cloth originated in Mesopotamia around 5000 BC"; "she measured off enough material for a dress")

 

Here is one example of one of its hypernyms:

 

·        computer network -- ((computer science) a network of computers)

o       network, electronic network -- ((electronics) a system of interconnected electronic components or circuits)

§         system -- (instrumentality that combines interrelated interacting artifacts designed to work as a coherent entity; "he bought a new stereo system"; "the system consists of a motor and a small computer")

·        instrumentality, instrumentation -- (an artifact (or system of artifacts) that is instrumental in accomplishing some end)

o       artifact, artefact -- (a man-made object taken as a whole)

§         object, physical object -- (a tangible and visible entity; an entity that can cast a shadow; "it was full of rackets, balls and other objects")

·        entity -- (that which is perceived or known or inferred to have its own distinct existence (living or nonliving))

§         whole, whole thing, unit -- (an assemblage of parts that is regarded as a single entity; "how big is that part compared to the whole?"; "the team is a unit")

·        object, physical object -- (a tangible and visible entity; an entity that can cast a shadow; "it was full of rackets, balls and other objects")

o       entity -- (that which is perceived or known or inferred to have its own distinct existence (living or nonliving))

 

The other relationships, such as hyponyms and meroynms, show predictable results and are not listed here.

 

Here is an example of the senses of the verb "net":

 

·        net, sack, sack up, clear -- (make as a net profit; "The company cleared $1 million")

·        net, clear -- (yield as a net profit; "This sale netted me $1 million")

·        web, net -- (construct or form a web, as if by weaving)

·        net, nett -- (catch with a net; "net a fish")

 

 

I also checked the senses for the adjective "net":

 

·        net (vs. gross), nett -- (remaining after all deductions; "net profit")

·        final, last, net -- (conclusive in a process or progression; "the final answer"; "a last resort"; "the net result")

 

The WordNet is a semi-formal, general ontology.  Its domain is the area of linguistics and human language, but it can be applied to other domains in which natural language processing is required.  It is an ontology because it provides more than a hierarchical architecture.  Its relational structure is more akin to a web than a tree.  It is a semi-formal ontology because it contents are not perfectly machine-readable and machine-processable.  For example, the definitions of the words in the database are only understandable by humans, whereas the actual relationships between words are understandable by the computer.  WordNet is useful for the Semantic Web because it allows computers to "understand" the structure of human language, which is an important component in the universe of human knowledge.  WordNet is used in applications such as web navigation, image retrieval, training systems, and multi-lingual document analysis.  WordNet will always be limited, however, by the nuances and complexities of language, which are bound to social and cultural notions and are very difficult to model.  Therefore WordNet is restricted by its formal relationship schema, but is nonetheless a powerful tool.