Statistically-representative sample data generation
Systems and methods for statistically-representative sample data generation are disclosed. For example, a sample-data generator and/or a data discriminator may be received by a system, which may utilize the sample-data generator to generate sample data. The data discriminator may be utilized to train the sample-data generator until the data discriminator cannot discriminate between data received from the sample-data generator and data received by a database associated with the system. The trained sample-data generator may be sent to other systems, which may generate and utilize, such as for prediction model training, statistically-representative sample data generated by the trained sample-data generator.
1. 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:
sending, from a first component of a first system situated in a first geographic location to a second component of a second system situated in a second geographic location, first instructions to generate sample data from first data stored in association with the second system, the first geographic location being different then the second geographic location and the first component having a first component type that is a same type as a second component type of the second component;
sending, from the first system to the second system, second instructions to identify differences between the sample data and the first data, the second system further configured to train a model to minimize the differences;
receiving an indication that the model has been trained such that the differences are less than a threshold level of difference;
receiving, at the first system and from the second system, a trained model;
causing the trained model to generate second data that is statistically-representative of the first data;
generating a database associated with the first system;
storing the second data in a first portion of the database, the first portion of the database indicating that the second data is associated with the second system;
storing third data associated with the first system in a second portion of the database; and
generating an interface configured to provide selective access to the database or at least one of the first portion or the second portion.
2. The system of claim 1 , the operations further comprising:
generating a predictive model configured to determine at least one of an outcome or a probability of the outcome occurring;
causing the predictive model to be trained utilizing at least a portion of the second data and at least a portion of third data stored in association with the first system; and
determining the at least one of the outcome or the probability of the outcome occurring using the predictive model as trained.
3. The system of claim 2 , the operations further comprising sending, from the first system and to the second system, an instance of the predictive model, the instance of the predictive model configured to accept the first data.
4. The system of claim 1 , wherein the database comprises a first database, the operations further comprising:
receiving the sample data; and
determining whether the sample data was received from a sample-data generator or from a second database.
5. The system of claim 1 , wherein at least one of the sample data, the first data, or the second data is associated with health-related data.
6. The system of claim 5 , wherein generating the second data comprises generating a first portion of health-related data and omitting a second portion of health-related data.
7. A method, comprising:
sending, from a first component of a first system located at a first geographic location to a second component of a second system located at a second geographic location, first instructions to generate sample data from first data stored in association with the second system, the first geographic location being different then the second geographic location and the first component having a first component type that is a same type as a second component type of the second component;
sending, from the first system to the second system, second instructions to train a model;
receiving, at the first system and from the second system, a trained model;
causing the trained model to generate second data that is statistically-representative of the first data;
generating a database associated with the first system;
storing the second data in a first portion of the database, the first portion of the database indicating that the second data is associated with the second system;
storing third data associated with the first system in a second portion of the database; and
generating an interface configured to provide selective access to the database or at least one of the first portion or the second portion.
8. The method of claim 7 , further comprising:
generating a predictive model configured to determine at least one of an outcome or a probability of the outcome occurring;
causing the predictive model to be trained utilizing at least a portion of the second data and at least a portion of third data stored in association with the first system; and
determining the at least one of the outcome or the probability of the outcome occurring using the predictive model as trained.
9. The method of claim 8 , further comprising sending, from the first system and to the second system, an instance of the predictive model, the instance of the predictive model configured to accept the first data.
10. The method of claim 7 , further comprising:
generating a first predictive model configured to determine a probability of an outcome occurring, the first predictive model trained utilizing at least a portion of third data associated with the first system;
determining a first confidence value associated with a first probability of the outcome occurring using the first predictive model;
generating a second predictive model configured to determine the probability of the outcome occurring, the second predictive model trained utilizing at least a portion of the second data;
determining a second confidence value associated with a second probability of the outcome occurring using the second predictive model;
determining that the second confidence value is more favorable than the first confidence value; and
identifying the second system as a priority system based at least in part on the second confidence value being more favorable than the first confidence value.
11. The method of claim 7 , further comprising:
identifying a feature of the second data that is absent from third data associated with the first system;
generating a first predictive model configured to determine a probability of an outcome occurring, the first predictive model trained utilizing a portion of the second data excluding the feature;
determining a first confidence value associated with a first probability of the outcome occurring using the first predictive model;
generating a second predictive model configured to determine the probability of the outcome occurring, the second predictive model trained utilizing at least a portion of the second data including the feature;
determining a second confidence value associated with a second probability of the outcome occurring using the second predictive model;
determining that the second confidence value is more favorable than the first confidence value; and
identifying the feature as a priority feature based at least in part on the second confidence value being more favorable than the first confidence value.
12. The method of claim 7 , wherein the trained model is a component of a generative adversarial network.
13. The method of claim 7 , wherein the trained model comprises a first trained model, and the method further comprises:
receiving, at the first system and from a third system, a second trained;
causing the second trained model to generate third data that is statistically-representative of fourth data associated with the third system; and
sending, to the third system, an interface configured to permit access to the second data and the third data.
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:
sending, from a first component of a first system located at a first geographic location to a second component of a second system located at a second geographic location, first instructions to generate sample data from first data stored in association with the second system, the first geographic location being different then the second geographic location and the first component having a first component type that is a same type as a second component type of the second component;
sending, from the first system to the second system, second instructions to train a model;
receiving, at the first system and from the second system, a trained model;
causing the trained model to generate second data that is statistically-representative of the first data;
generating a database associated with the first system;
storing the second data in a first portion of the database, the first portion of the database indicating that the second data is associated with the second system;
storing third data associated with the first system in a second portion of the database; and
generating an interface configured to provide selective access to the database or at least one of the first portion or the second portion.
15. The system of claim 14 , the operations further comprising:
generating a predictive model configured to determine at least one of an outcome or a probability of the outcome occurring;
causing the predictive model to be trained utilizing at least a portion of the second data and at least a portion of third data stored in association with the first system; and
determining the at least one of the outcome or the probability of the outcome occurring using the predictive model as trained.
16. The system of claim 15 , the operation further comprising sending, from the first system and to the second system, an instance of the predictive model, the instance of the predictive model configured to accept the first data.
17. The system of claim 14 , the operations further comprising:
generating a first predictive model configured to determine a probability of an outcome occurring, the first predictive model trained utilizing at least a portion of third data associated with the first system;
determining a first confidence value associated with a first probability of the outcome occurring using the first predictive model;
generating a second predictive model configured to determine the probability of the outcome occurring, the second predictive model trained utilizing at least a portion of the second data;
determining a second confidence value associated with a second probability of the outcome occurring using the second predictive model;
determining that the second confidence value is more favorable than the first confidence value; and
identifying the second system as a priority system based at least in part on the second confidence value being more favorable than the first confidence value.
18. The system of claim 14 , the operations further comprising:
identifying a feature of the second data that is absent from third data associated with the first system;
generating a first predictive model configured to determine a probability of an outcome occurring, the first predictive model trained utilizing a portion of the second data excluding the feature;
determining a first confidence value associated with a first probability of the outcome occurring using the first predictive model;
generating a second predictive model configured to determine the probability of the outcome occurring, the second predictive model trained utilizing at least a portion of the second data including the feature;
determining a second confidence value associated with a second probability of the outcome occurring using the second predictive model;
determining that the second confidence value is more favorable than the first confidence value; and
identifying the feature as a priority feature based at least in part on the second confidence value being more favorable than the first confidence value.
19. The system of claim 14 , wherein the trained model is a component of a generative adversarial network.
20. The system of claim 14 , wherein the trained model comprises a first trained model, and the operations further comprise:
receiving, at a first system and from a third system, a second trained model;
causing the second trained model to generate third data that is statistically-representative of fourth data associated with the third system; and
sending, to the third system, an interface configured to permit access to the second data and the third data.