Encoding multi-dimensional information for deep vision learning
AssigneeInternational Business Machines Corporation
InventorsShuyan Lu, Eugene Irving Kelton, Yi-Hui Ma, John H. Walczyk, III
Life of this patent
Estimating the term…
Abstract
A method, system, and computer program product for computer vision modeling are provided. The method identifies a set of transactions. A set of categorical behavior transaction types are determined for the set of transactions. The set of categorical behavior transaction types are mapped to a set of colors in a color coordinate system. The method scales color component values of the set of colors in the color coordinate system to generate a pattern of colorized units at intervals along a timespan of the set of transactions. The method generates a fraud detection model based on the set of transactions and the color component values.
Claims20
A computer-implemented method, comprising:
- identifying a set of transactions;
- determining a set of categorical behavior transaction types for the set of transactions;
- mapping the set of categorical behavior transaction types to a set of colors in a color coordinate system, each color of the set of colors in the color coordinate system having a set of color component values to form the color;
- scaling the set of transactions into color component values of the set of colors in the color coordinate system to generate a pattern of colorized units at intervals along a timespan of the set of transactions, the set of transactions being scaled by modifying one or more values associated with each transaction to fit at least one color component value of the set of color component values associated with a color mapped to a categorical behavior transaction type;
- and
- generating a fraud detection model based on the set of transactions and the color component values.
A system, comprising:
- one or more processors;
- and
- a non-transitory computer-readable storage medium, coupled to the one or more processors, storing program instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
- identifying a set of transactions;
- determining a set of categorical behavior transaction types for the set of transactions;
- mapping the set of categorical behavior transaction types to a set of colors in a color coordinate system, each color of the set of colors in the color coordinate system having a set of color component values to form the color;
- scaling the set of transactions into color component values of the set of colors in the color coordinate system to generate a pattern of colorized units at intervals along a timespan of the set of transactions, the set of transactions being scaled by modifying one or more values associated with each transaction to fit at least one color component value of the set of color component values associated with a color mapped to a categorical behavior transaction type;
- and
- generating a fraud detection model based on the set of transactions and the color component values.
A computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions being executable by one or more processors to cause the one or more processors to perform operations comprising:
- identifying a set of transactions;
- determining a set of categorical behavior transaction types for the set of transactions;
- mapping the set of categorical behavior transaction types to a set of colors in a color coordinate system, each color of the set of colors in the color coordinate system having a set of color component values to form the color;
- scaling the set of transactions into color component values of the set of colors in the color coordinate system to generate a pattern of colorized units at intervals along a timespan of the set of transactions, the set of transactions being scaled by modifying one or more values associated with each transaction to fit at least one color component value of the set of color component values associated with a color mapped to a categorical behavior transaction types;
- and
- generating a fraud detection model based on the set of transactions and the color component values.
Dependencies are inferred from the claim text.