Machine learning for medical screening recommendations based on patient activity information in social media

1. A method of analyzing patient activity to conduct appropriate medical tests comprising:

analyzing, via at least one processor, a collection of medical documents employing natural language processing to identify within document content risk factors of medical conditions associated with corresponding activities;

generating, via the at least one processor, feature sets of the identified risk factors for the medical conditions, wherein a feature is selected for a feature set in response to a feature score satisfying a threshold, and wherein the feature score is based on a distance within a medical ontology between the feature and a medical condition;

expanding, via the at least one processor, the feature sets with at least one feature similar to a corresponding feature of the feature sets, wherein the corresponding feature includes one from a group of hobby related activities and occupational related activities, and wherein the similarity of the at least one feature is based on a distance within a first ontology from the corresponding feature;

mapping, via the at least one processor, the identified risk factors of the activities to one or more medical tests for the associated medical conditions by training a machine learning model with the feature sets of the identified risk factors;

analyzing, via the at least one processor, at least one data source about a patient to determine patient activities and generate a feature set for the patient based on the determined patient activities, wherein the at least one data source includes a social media network indicating the patient activities including one or more selected from a group of hobby related activities and occupational related activities; and

determining, via the at least one processor, the one or more medical tests to conduct from the mapping based on the risk factors associated with the determined activities of the patient by:

determining a confidence score for the risk factors associated with the determined activities of the patient;

applying the feature set for the patient based on the determined patient activities and the confidence score for the risk factors associated with the determined activities of the patient to the machine learning model to provide the one or more medical tests to conduct;

validating the one or more medical tests provided by the machine learning model against prior performance for the patient, prior recommendation to the patient, and prior feedback received from a medical service provider for the patient; and

removing at least one medical test that failed the validation from the one or more medical tests provided by the machine learning model.

2. The method of claim 1 , wherein the identified risk factors are concepts within a second ontology, and mapping the identified risk factors comprises:

mapping concepts of the second ontology corresponding to the concepts of the identified risk factors to the one or more medical tests for the associated medical conditions.

3. The method of claim 1 , wherein the confidence score is based on one or more from a group of: concepts of an ontology present in a medical record and social media communications, a quantity of references to the concepts in the medical record and social media communications, and a distance within the ontology between the concepts and medical concepts identifying the medical tests.

4. The method of claim 1 , further comprising:

providing evidence to support a determination of the one or more medical tests to conduct.

5. The method of claim 4 , wherein the evidence includes one or more from a group of: the risk factors associated with the determined activities of the patient, and the confidence score for the risk factors associated with the determined activities of the patient.

6. The method of claim 1 , wherein the at least one data source further includes medical records associated with the patient.

7. A system for analyzing patient activity to conduct appropriate medical tests comprising:

at least one processor configured to:

analyze a collection of medical documents employing natural language processing to identify within document content risk factors of medical conditions associated with corresponding activities;

generate feature sets of the identified risk factors for the medical conditions, wherein a feature is selected for a feature set in response to a feature score satisfying a threshold, and wherein the feature score is based on a distance within a medical ontology between the feature and a medical condition;

expand the feature sets with at least one feature similar to a corresponding feature of the feature sets, wherein the corresponding feature includes one from a group of hobby related activities and occupational related activities, and wherein the similarity of the at least one feature is based on a distance within a first ontology from the corresponding feature;

map the identified risk factors of the activities to one or more medical tests for the associated medical conditions by training a machine learning model with the feature sets of the identified risk factors;

analyze at least one data source about a patient to determine patient activities and generate a feature set for the patient based on the determined patient activities, wherein the at least one data source includes a social media network indicating the patient activities including one or more selected from a group of hobby related activities and occupational related activities; and

determine the one or more medical tests to conduct from the mapping based on the risk factors associated with the determined activities of the patient by:

determining a confidence score for the risk factors associated with the determined activities of the patient;

applying the feature set for the patient based on the determined patient activities and the confidence score for the risk factors associated with the determined activities of the patient to the machine learning model to provide the one or more medical tests to conduct;

validating the one or more medical tests provided by the machine learning model against prior performance for the patient, prior recommendation to the patient, and prior feedback received from a medical service provider for the patient; and

removing at least one medical test that failed the validation from the one or more medical tests provided by the machine learning model.

8. The system of claim 7 , wherein the identified risk factors are concepts within a second ontology, and mapping the identified risk factors comprises:

mapping concepts of the second ontology corresponding to the concepts of the identified risk factors to the one or more medical tests for the associated medical conditions.

9. The system of claim 7 , wherein the confidence score is based on one or more from a group of: concepts of an ontology present in a medical record and social media communications, a quantity of references to the concepts in the medical record and social media communications, and a distance within the ontology between the concepts and medical concepts identifying the medical tests.

10. The system of claim 7 , wherein the at least one processor is further configured to:

provide evidence to support a determination of the one or more medical tests to conduct.

11. The system of claim 10 , wherein the evidence includes one or more from a group of: the risk factors associated with the determined activities of the patient, and the confidence score for the risk factors associated with the determined activities of the patient.

12. The system of claim 7 , wherein the at least one data source further includes medical records associated with the patient.

13. A computer program product for analyzing patient activity to conduct appropriate medical tests comprising:

one or more computer readable storage media collectively having program instructions embodied therewith, the program instructions executable by at least one processor to cause the at least one processor to:

analyze a collection of medical documents employing natural language processing to identify within document content risk factors of medical conditions associated with corresponding activities;

generate feature sets of the identified risk factors for the medical conditions, wherein a feature is selected for a feature set in response to a feature score satisfying a threshold, and wherein the feature score is based on a distance within a medical ontology between the feature and a medical condition;

expand the feature sets with at least one feature similar to a corresponding feature of the feature sets, wherein the corresponding feature includes one from a group of hobby related activities and occupational related activities, and wherein the similarity of the at least one feature is based on a distance within a first ontology from the corresponding feature;

map the identified risk factors of the activities to one or more medical tests for the associated medical conditions by training a machine learning model with the feature sets of the identified risk factors;

analyze at least one data source about a patient to determine patient activities and generate a feature set for the patient based on the determined activities, wherein the at least one data source includes a social media network indicating the patient activities including one or more selected from a group of hobby related activities and occupational related activities; and

determine the one or more medical tests to conduct from the mapping based on the risk factors associated with the determined activities of the patient by:

determining a confidence score for the risk factors associated with the determined activities of the patient;

applying the feature set for the patient based on the determined patient activities and the confidence score for the risk factors associated with the determined activities of the patient to the machine learning model to provide the one or more medical tests to conduct;

validating the one or more medical tests provided by the machine learning model against prior performance for the patient, prior recommendation to the patient, and prior feedback received from a medical service provider for the patient; and

removing at least one medical test that failed the validation from the one or more medical tests provided by the machine learning model.

14. The computer program product of claim 13 , wherein the identified risk factors are concepts within a second ontology, and mapping the identified risk factors comprises:

mapping concepts of the second ontology corresponding to the concepts of the identified risk factors to the one or more medical tests for the associated medical conditions.

15. The computer program product of claim 13 , wherein the confidence score is based on one or more from a group of: concepts of an ontology present in a medical record and social media communications, a quantity of references to the concepts in the medical record and social media communications, and a distance within the ontology between the concepts and medical concepts identifying the medical tests.

16. The computer program product of claim 13 , wherein the program instructions are further configured to cause the at least one processor to:

provide evidence to support a determination of the one or more medical tests to conduct.

17. The computer program product of claim 16 , wherein the evidence includes one or more from a group of: the risk factors associated with the determined activities of the patient, and the confidence score for the risk factors associated with the determined activities of the patient.