This agent is a semantic-based Natural Language Query interface that allows users to access data across multiple knowledge bases simultaneously (dbpedia and JPS). Semantic-based Natural Language Query is a method that converts human-readable natural languages such as English, Spanish, Chinese etc. into SPARQL or other equivalent semantic query languages in order to retrieve or verify information from a semantic database.
Such method usually leverages Natural Language Processing (NLP) tools using vast lexicons of trained language data and statistical analysis to break down grammatical structure of sentences. This agent uses the NLTK package and Stanford NLP tools. For different grammatical structure, the system will search for matching predefined SPARQL templates.
The lookup modules will subsequently convert recognised named-entities into URIs or IRIs and populate the SPARQL templates to form the SPARQL query. The user can either choose from the list of predefined queries or repeat the predefined queries in the query box with the keywords (bold) being modified. For example, London can be modified to Berlin or any other cities.
Natural Language Query
A semantic web based natural language query interface that allows users to access data across multiple knowledge bases simultaneously.(For example, dbpedia and JPS)
Among the predefined questions, the key words that have been highlighted can be changed. For example, in question: Show me all museums in London, London can be changed to Berlin or Beijing or other cities
In question: Show me all power plants in Singapore, Singapore can be changed to Other countries as well. e.g. Germany
However due to the limitation of the knowledge base, some query may not return answer.