System and method for employing model repository

A non-transitory computer readable storage medium having stored thereon computer executable instructions that, when executed on a computer, configure the compute to perform a method comprising:

• storing model metadata for each of a plurality of models, the model metadata comprising a statistical analysis technique identifier and one or more model input data identifiers;
• receiving a model execution request comprising data identifying one of the plurality of models, a model execution start date and a model execution end date;
• commencing, on the model execution start date, and ceasing on the model execution end date, execution of the model associated with the model execution request using a statistical analysis technique associated with the statistical analysis technique identifier and model input data associated with the one or more model input data identifiers;
• storing outputs of the executed model in a database, the outputs being associated with a model instance identifier, the model instance identifier uniquely identifying a model instance used to track the execution of the model and the outputs, the model instance comprising data describing a relationship between the model associated with the model execution request and an application of the model associated with the model execution request to the model input data, the model instance being associated with a model instance creation date and time, wherein most recent outputs of the executed model are stored in a layer of the database;
• creating a key associated with the model instance identifier and one or more of the model execution start date, an execution engine identifier, and a job identifier;
• and
• retrieving the outputs from the layer of the database using the key, by a consuming application.

A system comprising:

• memory operable to store at least one program;
• and
• at least one processor communicatively coupled to the memory, in which the at least one program, when executed by the at least one processor, causes the at least one processor to perform a method comprising the steps of:
• storing model metadata for each of a plurality of models, the model metadata comprising a statistical analysis technique identifier and one or more model input data identifiers;
• receiving a model execution request comprising data identifying one of the plurality of models, a model execution start date and a model execution end date;
• commencing, on the model execution start date, and ceasing on the model execution end date, execution of the model associated with the model execution request using a statistical analysis technique associated with the statistical analysis technique identifier and model input data associated with the one or more model input data identifiers;
• storing outputs of the executed model in a database, the outputs being associated with a model instance identifier, the model instance identifier uniquely identifying a model instance used to track the execution of the model and the outputs, the model instance comprising data describing a relationship between the model associated with the model execution request and an application of the model associated with the model execution request to the model input data, the model instance being associated with a model instance creation date and time, wherein most recent outputs of the executed model are stored in a layer of the database;
• creating a key associated with the model instance identifier and one or more of the model execution start date, an execution engine identifier, and a job identifier;
• and
• retrieving the outputs from the layer of the database using the key, by a consuming application.