Semantic hierarchical grouping of text fragments

1. A method, by a processor, for semantic hierarchical grouping of text fragments, comprising:

extracting sub-terms from a plurality of input text fragments according to a lexical sub-term hierarchy;

matching each of the sub-terms in the lexical sub-term hierarchy with concepts based on an ontology of concepts representing a domain knowledge; wherein the matching further includes assigning a concept score, at a first time, to each of the sub-terms matched with a respective concept, and identifying synonymous nodes between matching sub-terms within differing levels of the lexical sub-term hierarchy to create a virtual concept, wherein the virtual concept comprises the synonymous nodes which include each one of the sub-terms sharing an identical matching concept, the virtual concept including an alphanumeric concept identifier aggregated from each of a plurality of arbitrary alphanumeric concept identifiers, incorporating the concept score, associated with the respective sub-terms sharing the identical matching concept;

responsive to matching each of the sub-terms in the lexical sub-term hierarchy with the concepts and creating the virtual concept incorporating the sub-terms sharing the identical matching concept, re-scoring, at a second time, the concept score according to an average depth of the concepts within the lexical sub-term hierarchy; wherein, for each level of the lexical sub-term hierarchy, the re-scoring adds a predetermined value to an assigned root term token count and subtracts the predetermined value from a sub-term token count to determine the average depth;

automatically grouping the plurality of input text fragments into a hierarchy of concepts based on the matching, lexical sub-terms hierarchical relationships, and semantic relationships between matched concepts;

validating, by input from a user received through a user interface, an output of the hierarchy of concepts presented via the user interface;

upon receiving input from the user that the hierarchy of concepts in invalid, modifying the output of the hierarchy of concepts by one of: moving one or more nodes of the hierarchy of concepts, deleting one or more nodes of the hierarchy of concepts, or renaming one or more of the concepts; and

presenting instructions to the user via the user interface for using the output of the hierarchy of concepts for a clustering operation.

2. The method of claim 1 , wherein the matching further includes

selecting the concepts according to a highest concept score as compared to other concepts having a lower concept score for automatically grouping the plurality of input text fragments into the hierarchy of concepts.

3. The method of claim 1 , wherein the matching further includes matching semantic labels of the respective concept to one of the sub-terms using the ontology of concepts representing the domain knowledge.

4. The method of claim 1 , wherein the matching further includes filtering the concepts for the matching step according to defined filtering criteria, the defined filtering criteria including at least filtering the concepts belonging to a predefined set of semantic types or those concepts having a concept score above a defined threshold.

5. The method of claim 2 , further including

selecting the concepts according to a highest concept rescore as compared to other concepts having a lower concept rescore for automatically grouping the input text fragments into the hierarchy of concepts.

6. A system for assigning tasks to a robot device, comprising:

one or more computers with executable instructions that when executed cause the system to:

extract sub-terms from a plurality of input text fragments according to a lexical sub-term hierarchy;

match each of the sub-terms in the lexical sub-term hierarchy with concepts based on an ontology of concepts representing a domain knowledge; wherein the matching further includes assigning a concept score, at a first time, to each of the sub-terms matched with a respective concept, and identifying synonymous nodes between matching sub-terms within differing levels of the lexical sub-term hierarchy to create a virtual concept, wherein the virtual concept comprises the synonymous nodes which include each one of the sub-terms sharing an identical matching concept, the virtual concept including an alphanumeric concept identifier aggregated from each of a plurality of arbitrary alphanumeric concept identifiers, incorporating the concept score, associated with the respective sub-terms sharing the identical matching concept;

responsive to matching each of the sub-terms in the lexical sub-term hierarchy with the concepts and creating the virtual concept incorporating the sub-terms sharing the identical matching concept, re-score, at a second time, the concept score according to an average depth of the concepts within the lexical sub-term hierarchy; wherein, for each level of the lexical sub-term hierarchy, the re-scoring adds a predetermined value to an assigned root term token count and subtracts the predetermined value from a sub-term token count to determine the average depth;

automatically group the plurality of input text fragments into a hierarchy of concepts based on the matching, lexical sub-terms hierarchical relationships, and semantic relationships between matched concepts;

validate, by input from a user received through a user interface, an output of the hierarchy of concepts presented via the user interface;

upon receiving input from the user that the hierarchy of concepts in invalid, modify the output of the hierarchy of concepts by one of: moving one or more nodes of the hierarchy of concepts, deleting one or more nodes of the hierarchy of concepts, or renaming one or more of the concepts; and

present instructions to the user via the user interface for using the output of the hierarchy of concepts for a clustering operation.

7. The system of claim 6 , wherein the executable instructions

select the concepts according to a highest concept score as compared to other concepts having a lower concept score for automatically grouping the plurality of input text fragments into the hierarchy of concepts.

8. The system of claim 6 , wherein the matching further includes matching semantic labels of the respective concept to one of the sub-terms using the ontology of concepts representing the domain knowledge.

9. The system of claim 6 , wherein the matching further includes filtering the concepts for the matching step according to defined filtering criteria, the defined filtering criteria including at least filtering the concepts belonging to a predefined set of semantic types or those concepts having a concept score above a defined threshold.

10. The system of claim 7 , wherein the executable instructions

select the concepts according to a highest concept rescore as compared to other concepts having a lower concept rescore for automatically grouping the input text fragments into the hierarchy of concepts.

11. A computer program product for, by a processor, facilitating recording images to improve quality, the computer program product comprising a non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising:

an executable portion that extracts sub-terms from a plurality of input text fragments according to a lexical sub-term hierarchy;

an executable portion that matches each of the sub-terms in the lexical sub-term hierarchy with concepts based on an ontology of concepts representing a domain knowledge; wherein the matching further includes assigning a concept score, at a first time, to each of the sub-terms matched with a respective concept, and identifying synonymous nodes between matching sub-terms within differing levels of the lexical sub-term hierarchy to create a virtual concept, wherein the virtual concept comprises the synonymous nodes which include each one of the sub-terms sharing an identical matching concept, the virtual concept including an alphanumeric concept identifier aggregated from each of a plurality of arbitrary alphanumeric concept identifiers, incorporating the concept score, associated with the respective sub-terms sharing the identical matching concept;

an executable portion that, responsive to matching each of the sub-terms in the lexical sub-term hierarchy with the concepts and creating the virtual concept incorporating the sub-terms sharing the identical matching concept, re-scores, at a second time, the concept score according to an average depth of the concepts within the lexical sub-term hierarchy; wherein, for each level of the lexical sub-term hierarchy, the re-scoring adds a predetermined value to an assigned root term token count and subtracts the predetermined value from a sub-term token count to determine the average depth;

an executable portion that automatically groups the plurality of input text fragments into a hierarchy of concepts based on the matching, lexical sub-terms hierarchical relationships, and semantic relationships between matched concepts;

an executable portion that validates, by input from a user received through a user interface, an output of the hierarchy of concepts presented via the user interface;

an executable portion that, upon receiving input from the user that the hierarchy of concepts in invalid, modifies the output of the hierarchy of concepts by one of: moving one or more nodes of the hierarchy of concepts, deleting one or more nodes of the hierarchy of concepts, or renaming one or more of the concepts; and

an executable portion that presents instructions to the user via the user interface for using the output of the hierarchy of concepts for a clustering operation.

12. The computer program product of claim 11 , further including an executable portion that

selects the concepts according to a highest concept score as compared to other concepts having a lower concept score for automatically grouping the plurality of input text fragments into the hierarchy of concepts.

13. The computer program product of claim 11 , further including an executable portion that

matches semantic labels of the respective concept to one of the sub-terms using the ontology of concepts representing the domain knowledge.

14. The computer program product of claim 11 , further including an executable portion that filters the concepts for the matching step according to defined filtering criteria, the defined filtering criteria including at least filtering the concepts belonging to a predefined set of semantic types or those concepts having a concept score above a defined threshold.

15. The computer program product of claim 12 , further including an executable portion that

selects the concepts according to a highest concept rescore as compared to other concepts having a lower concept rescore for automatically grouping the input text fragments into the hierarchy of concepts.