Cognitive evaluation determined from social interactions

1. A computer-implemented method for determining cognitive issues, comprising:

converting, by a processor device, motions detected by body worn acceleration sensors and images captured by cameras of a security system into descriptions forming a patient's information data including social network data and daily living data, both relating to activities of daily living including social interactions, interaction durations, and interaction reply speeds by a patient, the patient's information data further including an emotional sentiment data associated with the social interactions;

training, by the processor device, a patient machine learning model on the patient's information data to generate a trained patient machine learning model, the trained patient machine learning model outputting a baseline for the patient, the baseline including the social interactions, the interaction durations, the interaction reply speeds, and the emotional sentiment data associated with the social interactions, the baseline being normal behavior and routines for the patient indicated by information data for the patient;

computing, by the processor device, changes between the baseline for the patient from the output of the trained patient machine learning model and new information data for the patient obtained from the body worn acceleration sensors and the cameras of the security system; and

evaluating, by the processor device, the changes between the baseline for the patient from the output of the trained patient machine learning model and the new information data for the patient in relation to the baseline to generate an indication as to whether the new information data for the patient is classified as normal or abnormal based on a probabilistic model.

2. The computer-implemented method as recited in claim 1 , wherein converting includes receiving social media feeds for the patient.

3. The computer-implemented method as recited in claim 1 , wherein converting includes receiving audio data and video data from an area around the patient.

4. The computer-implemented method as recited in claim 1 , further comprising generating a thumbprint for the patient and people known to the patient.

5. The computer-implemented method as recited in claim 4 , wherein generating the thumbprint includes utilizing physical characteristics of the patient for identifying the patient in the new social network data.

6. The computer-implemented method as recited in claim 1 , wherein evaluating includes cataloging and classifying emotions of the patient in the new social network data when interacting with both known people and unknown people.

7. The computer-implemented method as recited in claim 1 , further comprising sounding an alarm for a caregiver to intervene when the patient is interacting abnormally.

8. The computer-implemented method as recited in claim 1 , wherein evaluating includes developing patterns over time of interactions the patient has with known people.

9. The computer-implemented method as recited in claim 8 , wherein the patterns are selected from the group consisting of frequency of interaction, time of day of interaction, and day of week of interaction.

10. The computer-implemented method as recited in claim 1 , wherein evaluating includes analyzing sweat from the patient to detect a mood of the patient.

11. The computer-implemented method as recited in claim 1 , wherein the new information data for the patient is gathered from internet of things devices.

12. A computer program product for determining cognitive impairment, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform a method comprising:

converting, by a processor device, motions detected by body worn acceleration sensors and images captured by cameras of a security system into descriptions forming a patient's information data including social network data and daily living data both relating to activities of daily living including social interactions, interaction durations, and interaction reply speeds by a patient, the patient's information further including an emotional sentiment data associated with the social interactions;

training, by the processor device, a patient machine learning model on the patient's information data to generate a trained patient machine learning model, the trained patient machine learning model outputting a baseline for the patient, the baseline including the social interactions, the interaction durations, the interaction reply speeds, and the emotional sentiment data associated with the social interactions, the baseline being normal behavior and routines for the patient indicated by information data for the patient;

computing, by the processor device, changes between the baseline for the patient from the output of the trained patient machine learning model and new information data for the patient obtained from the body worn acceleration sensors and cameras of the security system; and

evaluating, by the processor device, the changes between the baseline for the patient from the output of the trained patient machine learning model and the new information data for the patient in relation to the baseline to generate an indication as to whether the new information data for the patient is classified as normal or abnormal based on probabilistic model.

13. A cognitive detection system for determining cognitive impairment, comprising:

a communication system connected to a communication network;

a processing system including a processor device and memory receiving a patient's information data, the processing system programmed to:

convert motions detected by body worn acceleration sensors and images captured by cameras of a security system into descriptions forming a patient's information data including social network data and daily living data, both relating to activities of daily living including social interactions, interaction durations, and interaction reply speeds by a patient, the patient's information data further including an emotional sentiment data associated with the social interactions;

train a patient machine learning model on the patient's information data to generate a trained patient machine learning model, the trained patient model outputting a baseline for the patient, the baseline including the social interactions between the patient and other people, the interaction durations, and the interaction reply speeds by the patient, the baseline being normal behavior and routines of the patient indicated by information data for the patient;

compute changes between the baseline for the patient from the output of the trained patient machine learning model and new information data for the patient obtained from the body worn acceleration sensors and cameras of the security system; and

evaluate the changes between the baseline for the patient from the output of the trained patient machine learning model and the new information data for the patient in relation to the baseline to generate an indication as to whether the now information data for the patient is classified as normal or abnormal based on a probabilistic model.

14. The system as recited in claim 13 , further comprising a wearable device that can analyze sweat of a wearer to determine mood.

15. The system as recited in claim 13 , further programmed to develop patterns over time of interactions the patient has with known people.

16. The system as recited in claim 15 , wherein the patterns are selected from the group consisting of frequency of interaction, time of day of interaction, and day of week of interaction.

17. The system as recited in claim 13 , further programmed to generate a thumbprint for the patient utilizing physical characteristics of the patient for identifying the patient in the new information data for the patient.

18. The system as cited in claim 13 , wherein the patient's information data include social media feeds for the patient.

19. The system as recited in claim 13 , further programmed to gather the new information data for the patient from internet of things devices.

20. The computer-implemented method as recited in claim 1 , wherein the emotional sentiment of the social interactions is selected from the group consisting of friendly, hostile, and benign.