Generating a classifier for performing a query to a given knowledge base

1. A computer device for generating a classifier for performing a query to a given knowledge base, the given knowledge base including a plurality of predicates, a plurality of subjects and a plurality objects related to each other, the computer device comprising:

a processor configured to:

select one predicate from the plurality of predicates;

generate, based on the given knowledge base, a plurality of triples, each triple of the plurality of triples including the one selected predicate, and a subject of the plurality of subjects and an object of the plurality of objects related to the one selected predicate;

generate a list of properties, each property of the list of properties being correlated to the subject and the object of the triple of the plurality of triples, the generation of the list of properties comprising performance of a context-based query within the given knowledge base;

generate a classifier having the list of properties related to the one selected predicate; and

transform a question or a portion of a document based on the generated classifier such that the question or the portion of the document includes a predefined predicate.

2. The computer device of claim 1 , wherein the processor is further configured to select subjects of the plurality of subjects and objects of the plurality of objects each having the one selected predicate.

3. The computer device of claim 1 , wherein the plurality of triples are RDF-based triples, and

wherein the subjects and objects of the plurality of triples are URI-based subjects and URI-based objects, respectively.

4. The computer device of claim 1 , wherein the processor is further configured to process a reference text of the given knowledge base to find at least one sentence with an occurrence of the subject and the object of one triple of the plurality of triples.

5. The computer device of claim 4 , wherein the processor is further configured, when processing the reference text, to generate a search query, the search query requesting sentences, in which the subject and the object of the one triple of the plurality of triples occur.

6. The computer device of claim 4 , wherein the processor is further configured to process the reference text for each triple of the plurality of triples.

7. The computer device of claim 4 , wherein the processor is further configured to convert subjects and objects occurring in requested sentences into generic subjects and generic objects.

8. The computer device of claim 7 , wherein the processor is further configured to generate the classifier based on a matrix, in which each of the requested sentences is represented as a feature vector.

9. The computer device of claim 1 , wherein the processor is further configured to:

select a number of predicates from the plurality of predicates;

generate a number of triples for each predicate of the number of predicates;

generate a list of properties for each of the number of triples; and

generate a classifier model including a plurality of classifiers having the list of properties related to each of the number of selected predicates.

10. The computer device of claim 9 , wherein the processor is further configured to generate the classifier model, the generation of the classifier model comprising use of a support vector machine.

11. The computer device of claim 10 , wherein the processor is further configured to use the support vector machine with the generated matrix.

12. The computer device of claim 1 , wherein the processor is further configured to receive a natural language query and to process the natural language query using the generated classifier.

13. The computer device of claim 12 , wherein the processor is further configured to process the natural language query using the generated classifier by converting the natural language query into a feature vector using the generated matrix.

14. A method for generating a classifier for performing a query to a given knowledge base, the given knowledge base including a plurality of predicates, a plurality of subjects and a plurality objects being related to each other, the method comprising:

selecting one predicate from the plurality of predicates;

generating, based on the given knowledge base, a plurality of triples, each triple of the plurality of triples including the one selected predicate and a subject of the plurality of subjects and an object of the plurality of objects related to the one selected predicate;

generating a list of properties, the generating of the list of properties comprising performing a context-based query within the given knowledge base, each property of the list of properties being correlated to the subject and the object of one triple of the plurality of triples;

generating a classifier having the list of properties related to the one selected predicate; and

transforming a question or a portion of a document based on the generated classifier such that the question or the portion of the document includes a predefined predicate.

15. The method of claim 14 , further comprising receiving a natural query and processing the natural language query using the generated classifier.