X hits on this document

PDF document

An LMF-based Web Service for Accessing WordNet-type Semantic Lexicons - page 6 / 7





6 / 7

Figure 7: Result of /pwd/word/synset/id/eng-09-08420278-n/?relation=hypo.

provides accesses to WordNet-type semantic lexicons, in- cluding PWN, WN-ja, and EDR. The service however is SOAP/WSDL-based, rather than REST. In addition, the data format adopted for representing the access results is not LMF-compliant.

7. Concluding Remarks

Figure 8: Alternative modeling Interlingual External Ref.




is not. They also proposed a schema for synset represen- tation that is employed by the Web service. The schema however is not LMF-compliant.

Recently, a language service type named concept dictio- nary has been published by the Language Grid, which

This paper presented a RESTful Web service for access- ing WordNet-type semantic lexicons. In particular, the API was described: we discussed the design of URI structure for querying, and outlined the output data format which is based on the Wordnet-LMF schema. Possible revisions to the Wordnet-LMF schema were also proposed to accom- modate innately bilingual/multilingual semantic lexicons like EDR. Other revisions to the XML schema of Wordnet- LMF include index attribute in the <Sense> element that allows the user to specify a sense number of a word form. These are however slight revisions and may not affect the generality of Wordnet-LMF as well as the original LMF. In fact, any XML document validated with the Wordnet- LMF schema is valid with our revised XML schema. For future work, inclusion of other wordnets should be con- sidered to demonstrate the applicability of the proposed framework. In addition, we will explore the issues asso- ciated with the notion of on-demand composite lexicon ac- cess. Such a Web service requires a mechanism to asso- ciate lexical entries across lexical resources on-demand/on- the-fly. To enable this, we have to develop an efficient Web service (hidden to users) to semantically align possi-


Document info
Document views7
Page views7
Page last viewedSat Oct 22 02:30:53 UTC 2016