Machine-learning-based predictive behavioral monitoring

Patent metadata

Field	Value
Publication number	US-12046372-B2
Application number	US-202016999133-A
Country	US
Kind code	B2
Filing date	Aug 21, 2020
Priority date	Aug 21, 2020
Publication date	Jul 23, 2024
Grant date	Jul 23, 2024

Abstract

Systems and methods are configured to perform machine-learning-based predictive behavioral response. In various embodiments, one or more behavioral monitoring data objects are identified and processed using a behavioral pattern prediction machine learning model to generate a behavioral pattern prediction model. The behavioral pattern prediction model is processed using a risk generation machine learning model to generate a risk model, wherein: (i) the risk generation machine learning model is generated based at least in part by one or more risk factors, and (ii) the risk model comprises a per-risk factor score for each risk factor of the one or more risk factors. The risk model is processed using an adjustment generation machine learning model to generate an adjustment model and one or more prediction-based actions are performed based on the adjustment model.

First claim

The invention claimed is:

1. A computer-implemented method comprising:
receiving, by one or more processors, one or more behavioral monitoring data objects generated by one or more distributed behavioral monitoring devices;
generating, by the one or more processors and using a behavioral pattern prediction model, one or more behavior data objects, wherein: (i) the one or more behavior data objects identify occurrences of an end user participating in one or more behaviors based at least in part on inferencing occurrences of the end user participating in the one or more behaviors, (ii) the behavioral pattern prediction model is generated by a behavioral pattern prediction machine learning model based at least in part on the one or more behavioral monitoring data objects, and (iii) the one or more behavior data objects are associated with one or more risk factors;
determining, by the one or more processors and using one or more risk models, one or more per-risk factor scores associated with progression towards respective ones of the one or more risk factors based at least in part on the one or more behavior data objects, wherein: (i) the one or more risk models are generated by a risk generation machine learning model based at least in part on an effect of the one or more behavior data objects on the one or more risk factors and (ii) the effect of the one or more behavior data objects on the one or more risk factors is based at least in part on one or more thresholds for respective ones of the one or more behavior data objects to affect the one or more risk factors;
determining, by the one or more processors and using an adjustment model, an adjustment to a contribution rate, wherein: (i) the adjustment model is generated by an adjustment generation machine learning model based at least in part on the one or more risk models and (ii) the adjustment is based at least in part on (a) a compensation associated with the progression towards the one or more risk factors, (b) one or more adjustment values associated with respective ones of the one or more per-risk factor scores, and (c) one or more weights associated with respective ones of the one or more per-risk factor scores, wherein the one or more weights are generated based at least in part on (1) a long-term effect and a short-term effect on the contribution rate associated with a respective one of the one or more risk factors, (2) a long-term model configured to optimize contributions over a first time period, and (3) a short-term model configured to optimize contributions over a second time period that is shorter than the first time period; and
initiating, by the one or more processors, the performance of one or more prediction-based actions based at least in part on the adjustment.

2. The computer-implemented method of claim 1 , wherein the risk generation machine learning model is configured to generate the one or more risk models by:
generating a general risk model based at least in part on the one or more behavior data objects;
generating an individual risk model based at least in part on the one or more behavior data objects in relation to the end user; and
combining the general risk model and the individual risk model.

3. The computer-implemented method of claim 1 , wherein the adjustment generation machine learning model comprises a time series prediction model and an adjustment application model.

4. The computer-implemented method of claim 3 , wherein the adjustment generation machine learning model generates the adjustment model by:
determining a projected contribution for a prospective period of time by using the time series prediction model; and
generating the adjustment model based at least in part on the projected contribution, the one or more risk models, and the adjustment application model.

5. The computer-implemented method of claim 1 , wherein the one or more prediction-based actions comprise automatically adjusting a current contribution rate to a financial instrument based at least in part on the adjustment model.

6. The computer-implemented method of claim 1 , wherein the one or more risk factors are determined based at least in part on one or more medical conditions applicable to the end user and the computer-implemented method further comprises:
identifying the one or more medical conditions based at least in part on a health profile data object for the end user and a medical conditions prediction machine learning model, wherein the medical conditions prediction machine learning model is configured to identify a probability for each medical condition of a plurality of medical conditions representative of a likelihood of the end user developing the medical condition.

7. The computer-implemented method of claim 5 , wherein the financial instrument is at least one of a health savings account (HSA) or a flexible spending account (FSA).

8. A system comprising memory and one or more processors configured to:
receive one or more behavioral monitoring data objects generated by one or more distributed behavioral monitoring devices;
generate, using a behavioral pattern prediction model, one or more behavior data objects, wherein: (i) the one or more behavior data objects identify occurrences of an end user participating in one or more behaviors based at least in part on inferencing occurrences of the end user participating in the one or more behaviors, (ii) the behavioral pattern prediction model is generated by a behavioral pattern prediction machine learning model based at least in part on the one or more behavioral monitoring data objects, and (iii) the one or more behavior data objects are associated with one or more risk factors;
determine, using one or more risk models, one or more per-risk factor scores associated with progression towards respective ones of the one or more risk factors based at least in part on the one or more behavior data objects, wherein: (i) the one or more risk models are generated by a risk generation machine learning model based at least in part on an effect of the one or more behavior data objects on the one or more risk factors and (ii) the effect of the one or more behavior data objects on the one or more risk factors is based at least in part on one or more thresholds for respective ones of the one or more behavior data objects to affect the one or more risk factors;
determine, using an adjustment model, an adjustment to a contribution rate, wherein: (i) the adjustment model is generated by an adjustment generation machine learning model based at least in part on the one or more risk models and (ii) the adjustment is based at least in part on (a) a compensation associated with the progression towards the one or more risk factors, (b) one or more adjustment values associated with respective ones of the one or more per-risk factor scores, and (c) one or more weights associated with respective ones of the one or more per-risk factor scores, wherein the one or more weights are generated based at least in part on (1) a long-term effect and a short-term effect on the contribution rate associated with a respective one of the one or more risk factors, (2) a long-term model configured to optimize contributions over a first time period, and (3) a short-term model configured to optimize contributions over a second time period that is shorter than the first time period; and
initiate the performance of one or more prediction-based actions based at least in part on the adjustment.

9. The system of claim 8 , wherein the risk generation machine learning model is configured to generate the one or more risk models by:
generating a general risk model based at least in part on the one or more behavior data objects;
generating an individual risk model base

Assignees

• Optum Inc

Inventors

• Muse Jon Kevin
• Boss Gregory J
• Godden Paul J
• Megerian Mark Gregory

Classifications

• G16H50/20

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title

• G16H50/70

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title

• G16H50/30 Primary

  for calculating health indices; for individual health risk assessment · CPC title