NEAREST NEIGHBOR PREDICTIONS FOR PROVIDING HEALTH INSIGHTS

• US 20170039335A1
• Filed: 12/02/2015
• Published: 02/09/2017
• Est. Priority Date: 08/06/2015
• Status: Active Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Ismail, Ansari Mohammed, Jassin, Ohad, Dumovic, Andrew Lindsay, Bitran, Hadas, Ronen, Royi, Yom-Tov, Elad, Albadawi, Haithem, Shariff, Farah, Holmdahl, Todd
• Application Number:
  US14/957,409
• Publication Number:
  US 20170039335A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G16H 40/63 for local operation
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

A network-accessible computer includes a network-communications interface, configured to receive health metrics of a user over a computer network. The network-accessible computer also includes a logic machine, which is configured to localize the user in a virtual space based on the health metrics, identify k nearest neighbors in the virtual space having k shortest Euclidean distances to the user, and generate a health insight comparing the user to the k nearest neighbors. The network-communications interface is further configured to send the health insight to a computing device associated with the user via the computer network.

Claims

1. A network-accessible computer, comprising:

a network-communications interface configured to receive health metrics of a user via a computer network;

a logic machine configured to localize the user in a virtual space based on the health metrics of the user;

the logic machine configured to identify k nearest neighbors in the virtual space having k shortest Euclidean distances to the user in the virtual space;

the logic machine configured to generate a health insight comparing the user to the k nearest neighbors; and

the network-communications interface configured to send the health insight to a computing device associated with the user via the computer network.

2. The network-accessible computer of claim 1, where the health metrics of the user include health metrics measured by one or more health sensors of one or more health-monitoring computing devices.

3. The network-accessible computer of claim 1, where the health metrics of the user include health metrics manually input via one or more input modalities of one or more computing devices.

4. The network-accessible computer of claim 1, where the virtual space includes a plurality of dimensions, and where each dimension of the plurality of dimensions corresponds to a different health metric of the health metrics of the user.

5. The network-accessible computer of claim 4, where each dimension of the plurality of dimensions is normalized to a shared scale.

6. The network-accessible computer of claim 4, where a scale of one or more of the dimensions is adjusted to a weighted scale which is different from scales of other dimensions.

7. The network-accessible computer of claim 4, wherein a localized position of the user in the virtual space is defined by a set of coordinates, and each coordinate in the set of coordinates corresponds to a different health metric.

8. The network accessible computer of claim 1, wherein the logic machine generates a health insight comparing the user to less than k nearest neighbors responsive to one or more of the k nearest neighbors having one or more health metrics significantly different from other users.

9. The network-accessible computer of claim 1, where the logic machine is configured to perform one or more statistical validation operations to confirm that the identified k nearest neighbors are suitable for comparison.

10. The network-accessible computer of claim 9, where the one or more statistical validation operations include:

determining that, for a given health metric, a difference between a value of the given health metric for the user and a mean value calculated for the k nearest neighbors is less than a first threshold fraction of a standard deviation of the given health metric calculated for a population of users; and

a standard deviation of the given health metric calculated for the k nearest neighbors is less than a second threshold fraction of the standard deviation of the given health metric calculated for the population of users.

11. The network-accessible computer of claim 10, where the first threshold fraction is equal to ⅓

, and the second threshold fraction is equal to ½

.

12. The network-accessible computer of claim 1, where the health insight indicates to the user how one or more health metrics of the user'"'"'s health metrics compare to one or more health metrics of the k nearest neighbors.

13. A network-accessible computer, comprising:

a network-communications interface configured to receive health metrics of a user from a health-monitoring computing device via a computer network, the user'"'"'s health metrics including a health metric measured by a health sensor of the health-monitoring computing device;

a logic machine configured to localize the user in a virtual space based on the user'"'"'s health metrics;

the logic machine configured to identify k nearest neighbors in the virtual space having k shortest Euclidean distances to the user;

the logic machine configured to generate a health insight comparing the user to the k nearest neighbors; and

the network-communications interface configured to send the health insight to the health-monitoring computing device associated with the user via the computer network.

14. The network-accessible computer of claim 13, where the virtual space includes a plurality of dimensions, and where each dimension of the plurality of dimensions corresponds to a different health metric of the health metrics of the user.

15. The network-accessible computer of claim 14, where each dimension of the plurality of dimensions is normalized to a shared scale.

16. The network-accessible computer of claim 14, where a scale of one or more of the dimensions is adjusted to a weighted scale which is different from scales of other dimensions.

17. The network-accessible computer of claim 14, wherein a localized position of the user in the virtual space is defined by a set of coordinates, and each coordinate in the set of coordinates corresponds to a different health metric.

18. The network-accessible computer of claim 13, where the logic machine is configured to perform one or more statistical validation operations to confirm that the identified k nearest neighbors are suitable for comparison, and the one or more statistical validation steps include:

determining that, for a given health metric, a difference between a value of the given health metric for the user and a mean value calculated for the k nearest neighbors is less than a first threshold fraction of a standard deviation of the given health metric calculated for a population of users; and

a standard deviation of the given health metric calculated for the k nearest neighbors is less than a second threshold fraction of the standard deviation of the given health metric calculated for the population of users.

19. A method for generating and presenting health insights, the method comprising:

receiving health metrics from a plurality of health-monitoring computing devices associated with a plurality of users;

localizing each particular user of the plurality of users in a virtual space according to the received health metrics;

for each particular user, identifying k nearest neighbors of the particular user by identifying k other users in the virtual space having k shortest Euclidean distances to the particular user;

for each particular user, generating one or more health insights by comparing one or more health metrics of the particular user'"'"'s health metrics to one or more health metrics associated with the k nearest neighbors; and

distributing generated health insights to the plurality of health-monitoring computing devices for presentation to the plurality of users.

20. The method of claim 19, where the virtual space includes a plurality of dimensions, and where each dimension of the plurality of dimensions corresponds to a different health metric of the health metrics;

and

wherein localizing each particular user in the virtual space comprises identifying, for each user, a set of coordinates, and each coordinate in the set of coordinates corresponds to a different health metric.