Question-Answering for Agricultural Open Data - Transactions on Large-Scale Data-and Knowledge-Centered Systems - page 24

Fig. 3. Interface of QA system

After searching the mappings registered in Google Big Table

and accessing the Microsoft Translator API and Japanese WordNet Ontology
provided by the National Institute of Information and Communications Tech-
nology (NICT), the LCS values for each mapping are calculated as described
in section 4.1 . The order of matching is firstly matching the Subject against
Resources by tracing 'sameAs' and 'wikiPageRedirects' links, and then search-
ing for Verb matches with the Properties of the Resources. A list of possible
answers is then created from the pairs of Properties and Values with the highest
LCS values. The number of answers in the list is set to three owing to constraints
on the client UI. The results of a Google search are also shown below in the client
to clarify the advantages and limitations of the QA system by comparison. The
user feedback is obtained by opening and closing a collapsible area in the client
that gives a detailed look at the Value of the Property. During searches, the
feedback updates the confidence value of a registered mapping

{ verb, property }

}
or registers a new mapping. During registration, the feedback has the role of
indicating which of three properties the object (value) should be registered to.
The client UI displays the results. Text-to-speech has not been implemented yet.
The query currently matches the graph pattern

{

verb, property

only. As the
target LOD, the QA system can search not only on the Plant Cultivation LOD,
but also on DBpedia.

<subject,verb, ? >