Does WordNet have “levels”? (NLP)

Question

For example...

Chicken animal .
Burrito is food .

WordNet allows you to execute the "is-a" ... hiearchy function.

However, how do I know when to stop traveling on a tree? I want LEVEL.
This is consistent.

For example, if you have a bunch of words, I want wordNet to classify all of them, but at a certain level, so it doesn't go too far. The categorization of “burrito” as “thing” is too broad, but “Mexican wrapped food” is too specific. I want to go up or down ... to the right LEVEL.

Answer 1

WordNet is a vocabulary, not an ontology, so "levels" really don't apply.

There is SUMO , which is the top ontology that applies to WordNet if you want a directional grid instead of a grid.

For some domains, the mid-level SUMO ontology is probably where you want to look, but I'm not sure that it has "Mexican wrapped food," since most of its topics are scientific or engineering.

WordNet Hierarchy

beef burrito < burrito < dish/2 < victuals < food < substance < entity. 

Essence is a higher-level concept, so if you stop a single substance, you get burritos and food. You can calculate the level based on this, but it will not necessarily be as consistent as SUMO, or it will create its own set of useful mid-level concepts to complete. WordNet does not have "Mexican packaged food."

Answer 2

[ Please forgive Pete Kirkham , he first came up with a link to SUMO, who may well answer the question posed by Alex, OP]

(I just provide more information here, I started in the comments field, but soon left the space and layout possibilities ...)

Alex : Most of SUMO is science or technology? It does not contain daily words such as food, people, cars, jobs, etc.

Pete K : SUMO is the top ontology. The mid-level ontologies (where you will find the concepts between “thing” and “beef burrito”) listed on the page do not include food, but reflect the types of organizations that finance the project. For people, there is a mid-level ontology. There is also one for industries (and therefore jobs), including food suppliers, but does not mention burritos if you learn it.

My two cents
100% WordNet (3.0, that is, the latest as well as earlier versions) is mapped to SUMO, and this may just be what Alex needs. Mid-level ontologies related to SUMO (or rather, MILO) work effectively in certain areas and currently do not include Foodstuff, but since WordNet does (including all - many, many of these everyday things), you don’t need to using some kind of formal ontology under "SUMO", but instead using Sumo WordNet mapping (perhaps in addition to WordNet, which, again, is not an ontology, but with its informal and free hierarchy) can also help.

However, some difficulties may arise from two areas (and then some ;-)?):

• The SUMO ontology level cannot be the level that you would have in mind for your particular application. For example, while “Burrito” brings “Food”, at the highest level, the essence in SUMO “Chicken” brings well “chicken”, which only finds “Animal” through a long chain (in particular: Chicken-> Poultry-> Bird- > Warm_Blooded_Vertebrae-> Vertebrae-> Animal).
• The coverage and metadata in Wordnet is impressive, but it can be a little inconsistent with mid-tier concepts. For example, “our” Burrito hypernym is appropriately “Dish,” which provides it with about 140 food items that include generics such as “Soup” or “Casserole”, as well as “Chicken Marengo” (but not saying “Chicken” Cacciatore ")

My point is not to criticize WordNet or SUMO and its related ontologies, but simply to illustrate simply some of the problems associated with building an ontology, especially at the intermediate level.

Regardless of some of the possible shortcomings and shortcomings of the solution based on SUMO and WordNet, the pragmatic use of these frameworks may well "match the score" (in 85% of cases)

Answer 3

To get levels, you need to predefine the content of each level. The ontology often defines them as the immediate children of an IS_A of a certain concept, but if this is missing, you need to develop a method for this yourself.

The next step is to prioritize each concept if you want to provide only one category for each word. Priority can be performed in several ways, for example, as the IS_A relationship between a category and a word, or manually selected priorities for each category. For each word, you can select the category with the highest priority. For example, you may want meat to be "food" rather than a chemical.

You can also select a few words that change priority if they are in the way. For example, if you want some chemicals that are also food to be declared chemicals, while others should still be food.

Answer 4

The WordNet hypernumeric tree ends with one syntactic root for the word entity. If you use the WordNet C library, then you can get the recursive structure for the synset ancestors with traceptrs_ds , and you can get the entire syntax tree recursively following the nextss and ptrlst until you click null pointers.

Answer 5

Sorry, can I ask what tool can evaluate the "level of complexity" of sentences? I want to find out a "similar level of difficulty" of suggestions for the user to read.