Generating a predictive model from multiple data sources
AssigneeInternational Business Machines Corporation
InventorsMarius I. Danciu, Fan Li, Michael McRoberts, Jing-Yun Shyr, Damir Spisic, Jing Xu
Life of this patent
Estimating the term…
Abstract
Techniques are disclosed for generating an ensemble model from multiple data sources. In one embodiment, the ensemble model is generated using a global validation sample, a global holdout sample and base models generated from the multiple data sources. An accuracy value may be determined for each base model, on the basis of the global validation dataset. The ensemble model may be generated from a subset of the base models, where the subset is selected on the basis of the determined accuracy values.
Claims15
1-7. (canceled)
A computer program product, comprising:
- a computer-readable storage medium having computer-readable program code embodied therewith, the computer-readable program code comprising:
- computer-readable program code configured to generate, for each of a plurality of data sources, a validation sample, a holdout sample, and a base model from the respective data source;
- computer-readable program code configured to merge the validation samples and the holdout samples for the plurality of data sources into a global validation dataset and a global holdout dataset, respectively;
- computer-readable program code configured to determine an accuracy value for each base model, based on the global validation dataset;
- and
- computer-readable program code configured to generate an ensemble model from a subset of the base models that are selected based on the determined accuracy values.
A system, comprising:
- one or more computer processors;
- a memory containing a program, which when executed by the one or more computer processors is configured to perform an operation comprising:
- for each of a plurality of data sources, generating a validation sample, a holdout sample, and a base model from the respective data source;
- merging the validation samples and the holdout samples for the plurality of data sources into a global validation dataset and a global holdout dataset, respectively;
- determining an accuracy value for each base model, based on the global validation dataset;
- and
- generating an ensemble model from a subset of the base models that are selected based on the determined accuracy values.
Dependencies are inferred from the claim text.