Mining new negation triggers dynamically based on structured and unstructured knowledge

A computer program product comprising a non-transitory computer readable storage medium having a computer readable program stored therein, wherein the computer readable program comprises instructions, which when executed on a processor of a computing device causes the computing device to implement a negation detection component within a cognitive natural language processing system, wherein the computer readable program causes the computing device to:

• analyze, by an attribute extraction component executing within the negation detection component, a portion of natural language text to identify an attribute specified in the natural language text, wherein the portion of natural language text comprises natural language text present in an electronic medical record of a patient and wherein the attribute is a medical attribute of the patient;
• analyze, by a compare component executing within the negation detection component, the portion of natural language text to determine whether a known negation trigger is present in the natural language text in association with the attribute;
• in response to determining that the natural language text does not contain a known negation trigger in association with the attribute, determine, by a sentence parsing component executing within the negation detection component, whether the attribute is negated based on instances of the attribute in other natural language content similar to the natural language text, comprising performing, by a clustering component executing within the negation detection component, cluster analysis to identify a cluster of entities having similar characteristics, wherein at least one entity in the cluster is associated with the natural language text, and wherein the other natural language content comprises a plurality of natural language content, each associated with a different entity in the cohort, wherein the entities of the cluster are other patients having at least one similar attribute to the patient and wherein the natural language content associated with the other patients in the cluster are electronic medical records of other patients;
• in response to determining that the attribute is negated, identify, by the sentence parsing component, a new negation trigger associated with the attribute in the natural language text;
• and
• store, by the negation detection component, the new negation trigger in association with the attribute in a negation trigger dictionary data structure.

A computing device comprising:

• a processor, and
• a memory coupled to the processor, wherein the memory comprises instructions, which when executed on a processor of a computing device causes the computing device to implement a negation detection component within a cognitive natural language processing system, wherein the instructions cause the processor to:
• analyze, by an attribute extraction component executing within the negation detection component, a portion of natural language text to identify an attribute specified in the natural language text, wherein the portion of natural language text comprises natural language text present in an electronic medical record of a patient and wherein the attribute is a medical attribute of the patient;
• analyze, by a compare component executing within the negation detection component, the portion of natural language text to determine whether a known negation trigger is present in the natural language text in association with the attribute;
• in response to determining that the natural language text does not contain a known negation trigger in association with the attribute, determine, by a sentence parsing component executing within the negation detection component, whether the attribute is negated based on instances of the attribute in other natural language content similar to the natural language text, comprising performing, by a clustering component executing within the negation detection component, cluster analysis to identify a cluster of entities having similar characteristics, wherein at least one entity in the cluster is associated with the natural language text, and wherein the other natural language content comprises a plurality of natural language content, each associated with a different entity in the cohort, wherein the entities of the cluster are other patients having at least one similar attribute to the patient and wherein the natural language content associated with the other patients in the cluster are electronic medical records of other patients;
• in response to determining that the attribute is negated, identify, by the sentence parsing component, a new negation trigger associated with the attribute in the natural language text;
• and
• store, by the negation detection component, the new negation trigger in association with the attribute in a negation trigger dictionary data structure.