CODEWITHSUNDEEP
javanlpwordnet
ramcrys
ramcrys

Reputation: 643

wordnet word phrase in synset

How can we find the word phrases in a synset ? In particular, take this synset for the adj "booked":

booked, engaged, set-aside -- (reserved in advance)

I use the RitaWN Java package (WordNet version is 2.1), and cannot seem to find the phrases. In the example above, when I run

RiWordnet wordnet = new RiWordnet(null);
String[] syn = wordnet.getSynset(word, "a", true);
for(int i = 0; i < syn.length; i++)
            System.out.println(syn[i]);

It only outputs

booked engaged

While "set-aside" is not listed.

I have tested a lot and all phrases are not found. Another example:

commodity, trade good, good -- (articles of commerce)

then "trade good" is not returned from the getSynset() method. So how can we actually get phrases ?

(the ritawn package is obtained from http://rednoise.org/rita/wordnet/documentation/index.htm)

Upvotes: 3

Views: 1718

Answers (2)

rednoyz
rednoyz

Reputation: 1327

RiTaWN seems to ignore "compound-words" by default. You can disable this to get the full list of phrases (line 2 below). 

RiWordnet wordnet = new RiWordnet();
wordnet.ignoreCompoundWords(false);
String[] syn = wordnet.getSynset("booked", "a", true);
System.out.println(Arrays.asList(syn));

Result:

[INFO] RiTa.WordNet.version [033]
[booked, engaged, set-aside]

Upvotes: 1

PBelzile
PBelzile

Reputation: 118

This answer is a bit off right field but in any case...

Idilia has an online Wordnet-like database that is actually much more complete and richer than Wordnet. Depending on where you are in your application it may make sense so I'm mentioning it. There are coding examples for Java access on the site.

In this case the query:

[{"fs":"booked/J1", "lemma":[], "definition":null}]

would return

{ "fs" : "booked/J1", "lemma" : [ "set_aside", "set-aside", "engaged", "booked" ], "definition" : "reserved in advance." }

Upvotes: 1

Related Questions