Machine learning-based system for estimating glucose values based on blood glucose measurements and contextual activity data

1. A processor-implemented method, comprising:

receiving a first set of inputs comprising contextual activity data collected from a source of user activity data, and a second set of inputs comprising intermittent glucose measurements provided from a blood glucose meter; and

processing, via an estimation model, the first set of inputs and the second set of inputs to generate a set of estimated real-time glucose values without using information from a continuous glucose monitor, wherein the generated set of estimated real-time glucose values are not inputted into the estimation model, and the estimation model is an ensemble model that comprises: one or more machine learning models, and a physiological model, wherein the physiological model is configured to be used in conjunction with the one or more machine learning models to confine the generated set of estimated real-time glucose values within an acceptable range.

2. The method of claim 1 , wherein the one or more machine learning models comprise at least one of:

one or more artificial neural networks;

one or more regression models;

one or more decision tree models;

one or more Bayesian networks;

one or more probabilistic machine learning models;

one or more Gaussian processing models; or

one or more Hidden Markov models.

3. The method of claim 1 , wherein the contextual activity data collected from the source of user activity data comprises one or more of:

metabolic data about a user; or

physical activity data about the user.

4. The method of claim 1 , wherein the receiving comprises:

receiving the first set of inputs comprising the contextual activity data collected from the source of user activity data, the second set of inputs comprising intermittent glucose measurements provided from the blood glucose meter, a third set of inputs comprising nutritional information input by a user, and a fourth set of inputs comprising insulin delivered by an insulin infusion device of the user; and

wherein the processing comprises:

processing, via the estimation model, the first set of inputs, the second set of inputs, the third set of inputs and the fourth set of inputs to generate the set of estimated real-time glucose values, wherein the set of estimated real-time glucose values track actual glucose values.

5. The method of claim 1 , wherein the processing is performed at a server system, wherein the estimation model is executed as part of a web application at the server system, and wherein the server system is configured to communicate results of generation of the set of estimated real-time glucose values to one or more of a client device and an insulin infusion device.

6. The method of claim 1 , wherein the processing is performed at:

a client device, and wherein the estimation model is executed as part of a mobile application at the client device.

7. The method of claim 1 , wherein the processing is performed at:

an insulin infusion device.

8. The method of claim 1 , further comprising:

generating a notification that responsive to determining that the set of estimated real-time glucose values are not within an acceptable accuracy range; or

controlling delivery of insulin by an insulin infusion device based on the set of estimated real-time glucose values responsive to determining that the set of estimated real-time glucose values are within the acceptable accuracy range.

9. The method of claim 1 , wherein the physiological model comprises a population-based model comprising one or more parameters weighted in accordance with one or more physiological parameters of a user.

10. A computing device comprising:

one or more processors; and

one or more processor-readable media storing instructions which, when executed by the one or more processors, cause performance of one or more operations comprising:

receiving a first set of inputs comprising contextual activity data collected from a source of user activity data, and a second set of inputs comprising intermittent glucose measurements provided from a blood glucose meter; and

processing, via an estimation model, the first set of inputs and the second set of inputs to generate a set of estimated real-time glucose values without using information from a continuous glucose monitor, wherein the generated set of estimated real-time glucose values are not inputted into the estimation model, and the estimation model is an ensemble model that comprises: one or more machine learning models; and a physiological model, wherein the physiological model is configured to be used in conjunction with the one or more machine learning models to confine the generated set of estimated real-time glucose values within an acceptable range.

11. The computing device of claim 10 , wherein the one or more machine learning models comprise at least one of:

one or more artificial neural networks;

one or more regression models;

one or more decision tree models;

one or more Bayesian networks;

one or more probabilistic machine learning models;

one or more Gaussian processing models; or

one or more Hidden Markov models.

12. The computing device of claim 10 , wherein the contextual activity data collected from the source of user activity data comprises one or more of:

metabolic data about a user; or

physical activity data about the user.

13. The computing device of claim 10 , the one or more operations further comprising:

generating a notification responsive to determining that the set of estimated real-time glucose values are not within an acceptable accuracy range; or

controlling delivery of insulin by an insulin infusion device based on the set of estimated real-time glucose values responsive to determining that the set of estimated real-time glucose values are within the acceptable accuracy range.

14. The computing device of claim 10 , wherein the receiving comprises:

receiving the first set of inputs comprising the contextual activity data collected from the source of user activity data, the second set of inputs comprising intermittent glucose measurements provided from the blood glucose meter, a third set of inputs comprising nutritional information input by a user, and a fourth set of inputs comprising insulin delivered by an insulin infusion device of the user; and

wherein the processing comprises:

processing, via the estimation model, the first set of inputs, the second set of inputs, the third set of inputs and the fourth set of inputs to generate the set of estimated real-time glucose values, wherein the set of estimated real-time glucose values track actual glucose values.

15. The computing device of claim 10 , wherein the physiological model comprises a population-based model comprising one or more parameters weighted in accordance with one or more physiological parameters of a user.

16. One or more non-transitory processor-readable media storing instructions which, when executed by one or more processors, cause performance of:

receiving a first set of inputs comprising contextual activity data collected from a source of user activity data, and a second set of inputs comprising intermittent glucose measurements provided from a blood glucose meter; and

processing, via an estimation model, the first set of inputs and the second set of inputs to generate a set of estimated real-time glucose values without using information from a continuous glucose monitor, wherein the generated set of estimated real-time glucose values are not inputted into the estimation model, and the estimation model is an ensemble model that comprises: one or more machine learning models; and a physiological model, wherein the physiological model is configured to be used in conjunction with the one or more machine learning models to confine the generated set of estimated real-time glucose values within an acceptable range.

17. The one or more non-transitory processor-readable media of claim 16 , wherein the one or more machine learning models comprise at least one of:

one or more artificial neural networks;

one or more regression models;

one or more decision tree models;

one or more Bayesian networks; one or more probabilistic machine learning models;

one or more Gaussian processing models; or

one or more Hidden Markov models.

18. The one or more non-transitory processor-readable media of claim 16 , wherein the receiving comprises:

receiving the first set of inputs comprising the contextual activity data collected from the source of user activity data, the second set of inputs comprising intermittent glucose measurements provided from the blood glucose meter, a third set of inputs comprising nutritional information input by a user, and a fourth set of inputs comprising insulin delivered by an insulin infusion device of the user; and

wherein the processing comprises:

processing, via the estimation model, the first set of inputs, the second set of inputs, the third set of inputs and the fourth set of inputs to generate the set of estimated real-time glucose values.

19. The one or more non-transitory processor-readable media of claim 16 , further storing instructions which, when executed by the one or more processors, cause performance of:

generating a notification responsive to determining that the set of estimated real-time glucose values are not within an acceptable accuracy range; or

controlling delivery of insulin by an insulin infusion device based on the set of estimated real-time glucose values responsive to determining that the set of estimated real-time glucose values are within the acceptable accuracy range.

20. The one or more non-transitory processor-readable media of claim 16 , wherein the physiological model comprises a population-based model comprising one or more parameters weighted in accordance with one or more physiological parameters of a user.