06.08.2015 Views

A Wordnet from the Ground Up

A Wordnet from the Ground Up - School of Information Technology ...

A Wordnet from the Ground Up - School of Information Technology ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

5.2. plWordNet at Three 175percentage of singleton synsets. It means that a large number of nominal lemmas aredescribed by several singleton synsets. The columns <strong>from</strong> 2nd to <strong>the</strong> 10th describepolysemous lemmas. Observe a slight tendency of verbal and adjectival lemmas tohave more meanings than do <strong>the</strong> nominal ones.Finally, Table 5.7 presents <strong>the</strong> number of instances of each lexico-semantic relation.Some relations are defined only for particular parts of speech. The table cells for <strong>the</strong>undefined combinations are filled with ‘—’. As expected, <strong>the</strong> derivational relationsdominate in <strong>the</strong> verbal and adjectival parts of plWordNet. For verbs, <strong>the</strong> set of definedrelations is clearly not rich enough. We plan to expand it in <strong>the</strong> future.RelationNo. instancesNouns Verbs Adjectives AllHypernymy 12150 687 155 12992Holonymy 1454 0 0 1454Meronymy 1563 0 0 1563Troponymy 0 37 0 37Antonymy 1212 173 1618 3003Conversion 35 66 0 101Relatedness 981 2618 1226 4825Pertainymy 1469 191 295 1955Fuzzynimy 640 44 423 1107Table 5.7: The number of lexico-semantic relations in plWordNetHypernymy among nouns seems to tend toward depth (perhaps due to <strong>the</strong> definitionof <strong>the</strong> synset, assumed in plWordNet), but <strong>the</strong> limited size of <strong>the</strong> plWordNet 1.0 doesnot allow fully justified conclusions yet. The longest hyponymy path has 11 links:{istota ‘being’}→{człowiek ‘human’, istota ludzka ‘human being’, homo sapiens‘Homo sapiens’, człowiek rozumny ‘sapient (human)’}→{człowiek ze względu na swoje zajęcie ‘human by occupation’}→{pracownik ‘employee’}→{pracownik ze względu na rodzaj pracy ‘employee by job’}→{pracownik instytucji publicznych ‘public servant’}→{funkcjonariusz ‘officer’}→{żołnierz ‘soldier’}→{komandos ‘commando, ranger’}→{spadochroniarz ‘paratrooper’, skoczek spadochronowy‘parachutists’}The semi-automatic expansion of plWordNet focused mainly on <strong>the</strong> nominal part.650 nominal synsets – a relatively high number – have been linked to more than one

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!