DATA ANALYTICS SYSTEM

• US 20140280142A1
• Filed: 03/14/2013
• Published: 09/18/2014
• Est. Priority Date: 03/14/2013
• Status: Active Grant

• Current Assignee: Leidos, Inc. (Leidos Holdings, Inc.), UChicago Argonne LLC (University of Chicago)
• Original Assignee: Science Applications International Corporation
• Inventors: Wasson, Glenn S., Covington, Susan, Silva, Adam
• Granted Patent: US 9,563,670 B2
• Time in Patent Office: Days
• US Class Current: 707/737
• CPC Class Codes: G06F 16/2465 Query processing support for facilitating data mining operations in structured databases
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

Methods, computer readable media, and apparatuses for building data models and performing model-based analytics are presented. A data analytics system may be implemented including software components to provide a data analytics platform and/or a data analytics programming interface. The components of the data analytics system may allow users to create and execute software applications to build data models model-based analytics functionality such as classification and prediction. For example, a programming interface may allow users to build data models of various predetermined model types, such as statistical data models, spatial statistical data models, graphical data models, pattern mining data models, clustering data models, and machine learning data models, among others, based on an input data set. A parallel computing infrastructure, for example, a cloud computing environment, may be used to build data models and perform model-based analytics with distributed processing and distributed storage.

Claims

1. One or more non-transitory computer-readable media storing computer-executable instructions which, when executed on a computer system, provide a software package comprising:

a model definition function configured to receive a model type parameter corresponding to one of a predetermined set of model types defined within the software package;

a plurality of model grouping functions, each said model grouping function configured to receive one or more data record grouping parameters, wherein each of the predetermined set of model types has a corresponding set of data record grouping parameters defined within the software package; and

a model creation function configured to create a data model by applying a set of model input data records to a model definition defined by the model type parameter received via the model definition function and one or more data record grouping parameters received via the plurality of model grouping functions.

2. The computer-readable media of claim 1, wherein the software package comprises at least one model grouping function configured to receive a data record field value parameter, at least one model grouping function configured to receive a time parameter, and at least one model grouping function configured to receive a location parameter.

3. The computer-readable media of claim 1, wherein the model creation function is configured to determine a set of computing resources in a parallel computing infrastructure, and use the determined set of computing resources to create the data model.

4. The computer-readable media of claim 3, wherein the model creation function is further configured to:

determine either a batch processing mode or a real time processing mode for creating the data model; and

use the set of computing resources to create the data model in accordance with either the batch processing mode or a real time processing mode.

5. The computer-readable media of claim 1, wherein the model creation function is configured to create a plurality of data models, wherein the number of data models created is based on the grouping parameters received via the plurality of model grouping functions.

6. The computer-readable media of claim 5, the software package further comprising a model analytics function configured to:

retrieve the plurality of data models;

apply an incoming data stream to each of the plurality of data models; and

classify the incoming data stream into one of the plurality of data models.

7. The computer-readable media of claim 1, the software package further comprising a plurality of model visualization functions, wherein each of the predetermined set of model types has a corresponding model visualization function.

8. A method, comprising:

receiving a model type parameter via a model definition function of a software package, the model type parameter corresponding to one of a predetermined set of model types;

providing a set of one or more grouping functions, wherein the grouping functions provided are based on the received model type parameter;

receiving one or more data record grouping parameters via the one or more grouping functions;

receiving a function call of a model creation function associated with the received model type parameter and the received data record grouping parameters; and

in response to the function call of a model creation function, creating a data model by applying a set of model input data records to a model definition defined by the model type parameter and the data record grouping parameters.

9. The method of claim 8, wherein the data record grouping parameters received via the one or more grouping functions comprises at least one of a data record field value parameter, a time parameter, or a location parameter.

10. The method of claim 8, wherein creating the data model comprises:

determining a set of computing resources in a parallel computing infrastructure for creating the data model.

11. The method of claim 10, wherein creating the data model further comprises:

determining either a batch processing mode or a real time processing mode for creating the data model; and

using the set of computing resources to create the data model in accordance with either the batch processing mode or a real time processing mode.

12. The method of claim 8, further comprising:

determining a plurality of model definitions based on the model type parameter and the data record grouping parameters; and

creating a plurality of data models corresponding to the data model groups, by applying the set of model input data records to each of the plurality of data model definitions.

13. The method of claim 12, further comprising:

retrieving the plurality of data models;

applying an incoming data stream to each of the plurality of data models; and

classifying the incoming data stream into one of the plurality of data models.

14. The method of claim 8, further comprising:

providing a set of one or more model visualization functions, wherein the model visualization functions provided are based on the received model type parameter.

15. One or more non-transitory computer-readable media storing computer-executable instructions which, when executed on a computer system, cause the computer system to:

invoke a model definition function, passing into said model definition function a model type parameter corresponding to one of a predetermined set of model types accepted by the model definition function;

invoke one or more from a set of available model grouping functions, passing into said model grouping functions one or more data record grouping parameters, wherein the set of available model grouping functions is based on the model type parameter passed into the model definition function; and

invoke a model creation function, said model creation function configured to create a data model by applying a set of model input data records to a model definition defined by the model type parameter passed into the model definition function and the one or more data record grouping parameters passed into the set of available model grouping functions.

16. The computer-readable media of claim 15, wherein the one or more invoked model grouping functions comprises one or more of a model grouping function configured to receive a data record field value parameter, a model grouping function configured to receive a time parameter, or a model grouping function configured to receive a location parameter.

17. The computer-readable media of claim 15, wherein the model creation function is configured to determine a set of computing resources in a parallel computing infrastructure, and use the determined set of computing resources to create the data model.

18. The computer-readable media of claim 17, wherein invoking the model creation function comprises passing into the model creation function a processing mode parameter corresponding to either a batch processing mode or a real time processing mode for creating the data model.

19. The computer-readable media of claim 15, wherein the invoked model creation function is configured to create a plurality of data models, wherein the number of data models created is based on the one or more data record grouping parameters passed into the set of available model grouping functions.

20. The computer-readable media of claim 15, storing further computer-executable instructions which, when executed on a computer system, cause the computer system to:

invoke a model visualization function configured to provide a graphical visual display for the created data model, the graphical visual display provided by the model visualization function having a type based on the model type parameter passed into the model definition function.