Source PaperYearLineSentence
C10-2097 2010 21
In (Shamsfard, 2008) a semi automatic method has been used for developing a lexical ontology called FarsNet for Persian language
C10-2097 2010 57
As in (Shamsfard, 2008), if the English transla tion of a given Persian word has only one synset in PWN, then the Persian word is linked to this PWN synset directly, or if for a candidate synset at least two English translations belong to it, then Persian word is linked to this PWN synset
C10-2097 2010 27
The existence of a single common synset in fact implies the existence of a single common sense between the two words and therefore their Persian translations shall be connected to this synset (Shamsfard, 2008)
C10-2097 2010 15
There have been some other efforts to create a WordNet for Persian language (Famian, 2007; Rouhizadeh et al, 2008; Shamsfard, 2008) but there exists no Persian WordNet yet that covers all Persian words in dictionary and comparable with PWN
C10-2097 2010 17
In (Shamsfard, 2008) a semi au tomatic method is proposed in which for each Persian word a number of PWN synsets are sug gested by system in order to be supervised by a human annotator to select a relevant synset
C10-2097 2010 18
Based on experiments mentioned by Shamsfard(2008), the proposed WordNet extracted automatically by the system, retrieved about 70% accura cy.In this paper a fully automatic method for con structing a large-scale Persian WordNet from available resource such as PWN, MRDs and corpora has been proposed
C10-2097 2010 116
The weighted average precision of the whole links in our automatically generated Persian WordNet is 82.6%, which outperforms the only comparable semi-automated Persian WordNet which was previously presented by (Shamsfard, 2008), about 12.5%