Predictive model generation
First Claim
1. A machine-based method comprising:
- in connection with a project in which a predictive model is generated based on historical data about a system being modeledselecting variables having at least a first predetermined level of significance from a pool of potential predictor variables associated with the historical data, to form a first population of predictor variables,extending the first population of predictor variables to include cross products of at least two variables, each being from the first population of predictor variables,selecting variables having at least a second predetermined level of significance from the extended first population of predictor variables to form a second population of predictor variables,extending the second population of predictor variables to include cross products of at least two variables, at least one of the variables for at least one of the cross products being from the pool of potential predictor variables that are associated with the historical data and having less than the first predetermined level of significance,selecting variables having at least a third predetermined level of significance from the extended second population of predictor variables to form a third population of predictor variables,automatically selecting a model generation method from a set of available model generation methods to match characteristics of the historical data,generating a possible model of the third population of predictor variables using a subsample of the historical data by the model generation method,determining whether the possible model generalizes to the historical data other than the subsample,applying the possible model to all of the historical data to generate a final model, cross-validating the final model using random portions of the historical data, and interacting with the system being modeled based on the final model.
12 Assignments
0 Petitions
Accused Products
Abstract
Models are generated using a variety of tools and features of a model generation platform. For example, in connection with a project in which a user generates a predictive model based on historical data about a system being modeled, the user is provided through a graphical user interface a structured sequence of model generation activities to be followed, the sequence including dimension reduction, model generation, model process validation, and model re-generation.
Historical multi-dimensional data is received representing multiple variables transformed to be maximally predictive for at least one outcome variable to be used as an input to a predictive model of a commercial system, model development process is validated for at one or more sets of such variables and enabling a user of a model generation tool to combine at least two of the variables from the sets of variables.
136 Citations
26 Claims
-
1. A machine-based method comprising:
-
in connection with a project in which a predictive model is generated based on historical data about a system being modeled selecting variables having at least a first predetermined level of significance from a pool of potential predictor variables associated with the historical data, to form a first population of predictor variables, extending the first population of predictor variables to include cross products of at least two variables, each being from the first population of predictor variables, selecting variables having at least a second predetermined level of significance from the extended first population of predictor variables to form a second population of predictor variables, extending the second population of predictor variables to include cross products of at least two variables, at least one of the variables for at least one of the cross products being from the pool of potential predictor variables that are associated with the historical data and having less than the first predetermined level of significance, selecting variables having at least a third predetermined level of significance from the extended second population of predictor variables to form a third population of predictor variables, automatically selecting a model generation method from a set of available model generation methods to match characteristics of the historical data, generating a possible model of the third population of predictor variables using a subsample of the historical data by the model generation method, determining whether the possible model generalizes to the historical data other than the subsample, applying the possible model to all of the historical data to generate a final model, cross-validating the final model using random portions of the historical data, and interacting with the system being modeled based on the final model. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26)
-
Specification