Systems and methods for modeling machine learning and data analytics

Abstract

Systems and methods for implementing and using a data modeling and machine learning lifecycle management platform that facilitates collaboration among data engineering, development and operations teams and provides capabilities to experiment using different models in a production environment to accelerate the innovation cycle. Stored computer instructions and processors instantiate various modules of the platform. The modules include a user interface, a collector module for accessing various data sources, a workflow module for processing data received from the data sources, a training module for executing stored computer instructions to train one or more data analytics models using the processed data, a predictor module for producing predictive datasets based on the data analytics models, and a challenger module for executing multi-sample hypothesis testing of the data analytics models.

1 . A system for executing data analytics, the system comprising:
 one or more processors; and   a memory coupled with the one or more processors, wherein the processor executes a plurality of modules stored in the memory, and wherein the plurality of modules comprises:
 a user interface for modeling and managing a data analytics plan, the user interface displaying a plurality of processing nodes, the processing nodes comprising one or more collector nodes, one or more workflow manager nodes, one or more training nodes, one or more predictor nodes and one or more challenger nodes; 
 a collector module, for accessing one or more data sources, each data source providing data for use in training and executing the data analytics plan, each instantiation of which is presented as one of the one or more collector nodes in the user interface; 
 a workflow module, for processing the data prior to its use in training and executing the data analytics plan, each instantiation of which is presented as one of the one or more workflow manager nodes in the user interface; 
 a training module, for training one or more data analytics models using the processed data, each instantiation of which is presented as one of the one or more training manager nodes in the user interface; 
 a predictor module, for producing one or more predictive datasets based at least in part on the one or more data analytics models, each instantiation of which is presented as one of the one or more predictor nodes in the user interface; and 
 a challenger module, for executing multi-sample hypothesis testing of the data analytics models, each instantiation of which is presented as one of the one or more challenger nodes in the user interface. 
   
     
     
         2 . The system of  claim 1  further comprising a publishing module, for publishing results calculated by the predictor module, wherein each instantiation of the publishing module is presented as a publisher node in the user interface. 
     
     
         3 . The system of  claim 1  wherein the one or more data sources comprise at least one static data source accessed using an application programming interface to an external computing system. 
     
     
         4 . The system of  claim 1  wherein the one or more data sources comprise at least one real-time data streaming source. 
     
     
         5 . The system of  claim 1  wherein the workflow module implements a series of tasks, each task performing a discrete operation on the data. 
     
     
         6 . The system of  claim 5  wherein the discrete operation is selected from a group comprising data filtering, data aggregation, data selection, data parsing and data normalization. 
     
     
         7 . The system of  claim 5  further comprising a subordinate user interface in which a directed acyclic graph is constructed comprising tasks represented by vertices in the directed acyclic graph and possible paths of the data represented as edges in the directed acyclic graph. 
     
     
         8 . The system of  claim 5  wherein each instance of the workflow module is assigned to a specified processor such that the series tasks associated with the instance of the workflow module are executed on the specified processor. 
     
     
         9 . The system of  claim 8  wherein the specified processor is selected automatically based at least in part on an estimated data processing load of the series of tasks. 
     
     
         10 . The system of  claim 1  wherein the multi-sample hypothesis testing comprises executing two or more of the data analytics models. 
     
     
         11 . The system of  claim 10  wherein executing two or more of the data analytics models comprises allocating a first subset of the data to one of the two or more models, and a second subset of the data to a second of the two or more models. 
     
     
         12 . The system of  claim 11  wherein the allocation of the data to the two or more models comprises allocating a percentage of the data to each of the two or more models. 
     
     
         13 . The system of  claim 11  wherein the allocation of the data to the two or more models comprises allocating substantially all of the data to a first model and, upon receiving an indication the first model is not performing, sending substantially all of the data to a second model. 
     
     
         14 . The system of  claim 11  wherein the allocation of the data to the two or more models comprises allocating substantially all of the data to each model in parallel and designating a priority for at least one of the two or more models. 
     
     
         15 . A method for implementing data analytics models, the method comprising:
 displaying, on a user interface for modeling and managing a data analytics plan, a plurality of processing nodes, the processing nodes comprising one or more collector nodes, one or more workflow manager nodes, one or more training nodes, one or more predictor nodes and one or more challenger nodes;   accessing one or more data sources, each data source providing data for use in training and executing the data analytics plan, and presenting each data source as one of the one or more collector nodes in the user interface;   processing the data prior to its use in training and executing the data analytics plan, and presenting each processing step as one of the one or more workflow manager node in the user interface;   training one or more data analytics models using the processed data, and presenting each training step as one of the one or more training manager nodes in the user interface;   producing one or more predictive datasets based at least in part on the one or more data analytics models, and presenting each production step as one of the one or more predictor nodes in the user interface; and   executing multi-sample hypothesis testing of the data analytics models, and presenting each testing step as one of the one or more challenger nodes in the user interface.   
     
     
         16 . The method of  claim 15  further comprising publishing the predictive datasets, and presenting each publishing step as a publisher node in the user interface. 
     
     
         17 . The method of  claim 15  wherein the one or more data sources comprise at least one static data source accessed using an application programming interface to an external computing system. 
     
     
         18 . The method of  claim 15  wherein the one or more data sources comprise at least one real-time data streaming source. 
     
     
         19 . The method of  claim 15  wherein processing the data prior to its use in training and executing the data analytics plan comprises a series of tasks, each task performing a discrete operation on the data. 
     
     
         20 . The method of  claim 19  wherein the discrete operation is selected from a group comprising data filtering, data aggregation, data selection, data parsing and data normalization. 
     
     
         21 . The method of  claim 19  further comprising displaying a subordinate user interface in which a directed acyclic graph is constructed comprising tasks represented by vertices in the directed acyclic graph and possible paths of the data represented as edges directed acyclic graph. 
     
     
         22 . The method of  claim 19  wherein the series of tasks are assigned to a specified processor such that the series tasks associated with the instance of the workflow module are executed on the specified processor. 
     
     
         23 . The method of  claim 22  wherein the specified processor is assigned by a human. 
     
     
         24 . The method of  claim 22  wherein the specified processor is selected automatically based at least in part on an estimated data processing load of the series of tasks. 
     
     
         25 . The method of  claim 15  wherein the multi-sample hypothesis testing comprises executing two or more of the data analytics models. 
     
     
         26 . The method of  claim 25  wherein executing two or more of the data analytics models comprises allocating a first subset of the data to one of the two or more models, and a second subset of the data to a second of the two or more models. 
     
     
         27 . The method of  claim 26  wherein the allocation of the data to the two or more models comprises allocating a percentage of the data to each of the two or more models. 
     
     
         28 . The method of  claim 26  wherein the allocation of the data to the two or more models comprises allocating substantially all of the data to a first model and, upon receiving an indication the first model is not performing, sending substantially all of the data to a second model. 
     
     
         29 . The method of  claim 26  wherein the allocation of the data to the two or more models comprises allocating substantially all of the data to each model in parallel and designating a priority for at least one of the two or more models. 
     
     
         30 . An article of manufacture having computer-readable instructions stored thereon for:
 a user interface for modeling and managing a data analytics plan, the user interface displaying a plurality of processing nodes, the processing nodes comprising one or more collector nodes, one or more workflow manager nodes, one or more training nodes, one or more predictor nodes and one or more challenger nodes;   a collector module, for accessing one or more data sources, each data source providing data for use in training and executing the data analytics plan, each instantiation of which is presented as one of the one or more collector nodes in the user interface;   a workflow module, for processing the data prior to its use in training and executing the data analytics plan, each instantiation of which is presented as one of the one or more workflow manager nodes in the user interface;   a training module, for training one or more data analytics models using the processed data, each instantiation of which is presented as one of the one or more training manager nodes in the user interface;   a predictor module, for producing one or more predictive datasets based at least in part on the one or more data analytics models, each instantiation of which is presented as one of the one or more predictor nodes in the user interface; and   a challenger module, for executing multi-sample hypothesis testing of the data analytics models, each instantiation of which is presented as one of the one or more challenger nodes in the user interface.