Segmented actuarial modeling

1. A system, comprising:

one or more processors; and

computer-readable media storing computer-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:

receiving claims data indicating a claim status, the claim status including at least one of a pending status, a paid status, or a denied status;

generating metrics based at least in part on the claims data, the metrics including:

a first metric including a first number of the claims from the claims data;

a second metric including a second number of the claims indicated to be paid;

a third metric including price data representing a contractual price for individual ones of the second number of the claims; and

a fourth metric including a plan responsibility value associated with the second number of the claims;

generating training data by using a portion of the claims data associated with the metrics;

generating additional training data by resampling the portion of the claims data based at least in part on the training data including insufficient data for accurate predictions, wherein the insufficient data for accurate predictions is based at least in part on a confidence level associated with the training data;

training one or more machine learning (ML) models using the training data including the additional training data to classify input metrics by metrics categories and to predict output reserves with associated confidence levels and determine associated weights for the metrics, wherein training the one or more ML models includes generating an actuarial model to receive the input metrics and to determine the output reserve with an associated confidence level that is above threshold level;

determining a reserve prediction value using the actuarial model;

storing the metrics and the reserve prediction value as data objects, the data objects including a label with a current date;

causing an application to change an amount of money held in reserve based at least in part on the reserve prediction value, wherein the change to the amount of money is performed automatically by the system;

generating a user interface configured to display the metrics and the reserve prediction value, the user interface including one or more selectable options to change one or more of the metrics; and

causing the user interface to display an alert associated with the reserve prediction value based at least in part on the change to the amount of money being performed automatically.

2. The system of claim 1 , wherein the application comprises a first application, and the operations further comprise:

determining a change threshold based at least in part on a pattern of incremental changes for the claim status of the claims being denied during a specified time period;

determining that a change in a denial rate exceeds the change threshold based at least in part on the first metric and the second metric;

generating an alert based at least in part on the change in the denial rate exceeding the change threshold; and

sending notification data representing the alert to a computing device, the notification data causing a second application associated with the computing device to initiate and display at least a portion of the alert.

3. The system of claim 2 , the operations further comprising:

generating a time series graph associated with the denial rate within the specified time period; and

indicating on the time series graph the change in the denial rate exceeding the change threshold.

4. The system of claim 1 , wherein the reserve prediction value comprises a first reserve prediction value, and the operations further comprise:

receiving an indication that the contractual price has increased or decreased;

causing the third metric to be updated to an updated third metric based at least in part on the indication;

determining, using the actuarial model, a second reserve prediction value based at least in part on the updated third metric; and

causing the application to change the amount of money held in reserve based at least in part on the second reserve prediction value.

5. The system of claim 1 , wherein the reserve prediction value comprises a first reserve prediction value, and the operations further comprise:

causing the user interface to display the first metric, the second metric, the third metric, and the fourth metric as editable fields;

receiving input data representing user input to change at least one of the first metric, the second metric, the third metric, or the fourth metric;

determining a second reserve prediction value based at least in part on the input data; and

causing the user interface to display the second reserve prediction value.

6. A method, comprising:

receiving data associated with claims;

determining metrics based at least in part on the data, the metrics including:

a first metric including a first number of the claims from the data;

a second metric including a second number of the claims indicated to be paid; and

a third metric including price data representing a contractual price for individual ones of the second number of the claims;

generating training data by using a portion of the data associated with the metrics;

training one or more machine learning (ML) models using the training data to classify input by metrics categories and to predict an output reserve, wherein the one or more ML models includes an actuarial model to receive input metrics and to determine the output reserve;

determining a reserve prediction value using the actuarial model;

storing the metrics and the reserve prediction value as data objects, the data objects including a label with a current date; and

causing an application to change an amount of money held in reserve based at least in part on the reserve prediction value.

7. The method of claim 6 , further comprising:

determining the training data includes insufficient data for accurate predictions, wherein to make accurate predictions is based at least in part on determining a confidence level associated with the training data;

generating additional data for the training data by resampling the portion of the data associated with the metrics, wherein second training data includes the training data and the additional data;

determining the second training data includes sufficient data for accurate predictions, wherein the accurate predictions is based at least in part on a confidence level associated with the second training data is above a threshold level; and

generating a user interface configured to display the metrics and the reserve prediction value, the user interface including one or more selectable options to change one or more metrics.

8. The method of claim 6 , wherein the data comprises first data and the metrics comprises first metrics, and further comprising:

receiving second data associated with the claims;

determining second metrics based at least in part on the second data, a fourth metric of the second metrics includes a plan responsibility value associated with the second number of the claims; and

training new ML models using the second data as new training data, the new ML models includes a second actuarial model.

9. The method of claim 6 , further comprising:

determining that a denial rate change exceeds a threshold based at least in part on the first metric and the second metric; and

generating an alert based at least in part on the denial rate change.

10. The method of claim 9 , wherein the alert identifies one or more factors affecting the denial rate change, the one or more factors including one or more of a benefit design change, a premium rate change, or an aging members block.

11. The method of claim 9 , wherein the application further comprises a first application, and further comprising:

generating a time series graph associated with the denial rate change; and

sending notification data representing the alert to a computing device, the notification data causing a second application associated with the computing device to initiate and display at least a portion of the alert and the time series graph.

12. The method of claim 6 , wherein the reserve prediction value comprises a first reserve prediction value, and further comprising:

causing a user interface to display at least one of the first metric, the second metric, or the third metric as editable fields;

receiving input data representing user input to change the at least one of the first metric, the second metric, or the third metric;

determining a second reserve prediction value based at least in part on the input data; and

causing the user interface to display the second reserve prediction value.

13. The method of claim 12 , further comprising:

receiving second input data representing second user input to save the change to the at least one of the first metric, the second metric, or the third metric; and

generating a second actuarial model based at least in part on saving the change.

14. A system, comprising:

one or more processors; and

computer-readable media storing first computer-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:

receiving data associated with claims;

determining one or more metrics from the data based at least in part on information correlated with predicting reserves for the claims;

generating training data by using a portion of the data associated with the one or more metrics;

generating additional training data by resampling the portion of the data based at least in part on the training data including insufficient data for accurate predictions, wherein the insufficient data for accurate predictions is based at least in part on a confidence level associated with the training data;

training an actuarial model using the training data and the additional training data to receive input metrics and output a reserve prediction;

determining a reserve prediction value using the actuarial model;

storing the metrics and the reserve prediction value as data objects; and

causing an application to change an amount of money held in reserve based at least in part on the reserve prediction value.

15. The system of claim 14 , wherein generating the additional training data further comprises:

generating the additional training data by projecting claims data from the data associated with the claims and projecting corresponding metrics.

16. The system of claim 14 , the operations further comprising:

generating new training data by including the data objects with the training data; and

training a new actuarial model using the new training data.

17. The system of claim 14 , the operations further comprising:

generating a user interface configured to display the one or more metrics and the reserve prediction value, the user interface including one or more selectable options to change at least one of the one or more metrics.

18. The system of claim 14 , the operations further comprising:

generating a time series graph associated with the one or more metrics within a specified time period;

determining that a rate of change for at least one of the one or more metrics exceeds a threshold; and

indicating on the time series graph the rate of change for the at least one of the one or more metrics exceeding the threshold.

19. The system of claim 14 , the operations further comprising:

performing stochastic modeling on the one or more metrics;

determining a confidence level for the one or more metrics; and

determining to automatically perform data analysis daily based at least part on the confidence level being above a threshold.

20. The system of claim 14 , the operations further comprising:

performing stochastic modeling on the one or more metrics;

determining a confidence level for the one or more metrics;

changing a second metric of the one or more metrics based at least in part on the confidence level being below a threshold level;

storing the one or more metrics as data objects including labels for dates;

generating new training data by including the data objects with the training data;

training a second actuarial model using the new training data; and

determining a third prediction using the second actuarial model.