Automatic Generation of Synsets for Wordnet of Hindi Language

2. Conventional Way Of Automatic Wordnet Generation

Before heading towards our approach in constructing wordnets and translating the senses of each term for multiple languages, this section will introduce the conventional approaches or models in creating wordnets that was used by early researchers.

First approach is merge model, in which conversion of thesaurus into a wordnet is mentioned. Then, this wordnet is linked semi-automatically to another wordnet. The disadvantage is this model is only confined to small number of languages. Unless there are few available word-net for every language, it is not applicable to a wide span of languages.

Second approach is the expand model, is better than the previous one as it consumes less pre-existing resources. This uses a strategy as follows:

• 1.

  Consider an available wordnet for some language L_O, generally English Princeton WordNet;

• 2.

  In every sense s mentioned in the wordnet, then, using translation dictionary, translation of the elements linked through s from L_O to destination language L_N is taken place;

• 3.

  In addition, preserve all related relations concerned with semantics among senses from t already available wordnet so that it can arrive for a new wordnet i.e. L_N.