Systems and methods for generating transaction profile tags

1. A method, comprising:

associating, with at least one processor, at least one transaction label from among a plurality of transaction labels with each of a plurality of recorded transactions, each transaction label associated with a transaction type;

generating, with at least one processor, a set of pre-processed feature values corresponding to a set of profile features based on the plurality of recorded transactions from a transaction profile and the transaction labels corresponding to each of the plurality of recorded transactions, the set of profile features having a first dimension;

processing, with at least one processor, the set of pre-processed feature values with a macro-encoder to generate a first-reduced set of pre-processed feature values corresponding to a first-reduced set of profile features having a second dimension less than the first dimension;

clustering, with at least one processor, the first-reduced set of pre-processed feature values into at least two subsets of first-reduced feature values, a first subset of first-reduced feature values associated with a first macro-profile tag and a second subset of first-reduced feature values associated with a second macro-profile tag; and

tagging, with at least one processor, the transaction profile with either the first macro-profile tag or the second macro-profile tag based on the clustering of the first-reduced set of feature values,

wherein the first macro-profile tag is associated with transaction profiles having features corresponding to low-threshold purchases, and the second macro-profile tag is associated with transaction profiles having features corresponding to high-threshold purchases.

2. The method of claim 1 , further comprising:

determining, with at least one processor, an ordering of at least two subsets of first-reduced feature values.

3. The method of claim 1 , further comprising:

processing the set of pre-processed feature values with a micro-encoder to generate a second-reduced set of pre-processed feature values; and

clustering the second-reduced set of feature values into a plurality of subsets of second-reduced feature values, wherein each subset of the second-reduced feature values is associated with a micro-profile tag corresponding to a unique account profile type.

4. The method of claim 3 , further comprising:

determining, with at least one processor, an ordering of the plurality of subsets of second-reduced feature values.

5. The method according to claim 1 , wherein each account profile type from among a plurality of account profile types is associated with features corresponding to a partially-unique set of behavior characteristics.

6. The method according to claim 5 , further comprising:

receiving, with at least one processor, an instant transaction;

generating, with at least one processor, at least one feature value for the instant transaction;

determining, with at least one processor, that the at least one feature value does not correspond to a feature associated with the account profile type; and

determining, with at least one processor, that the instant transaction is a suspect transaction based on the determination that the at least one feature value does not correspond to a feature associated with the account profile type.

7. The method according to claim 5 , further comprising:

receiving, with at least one processor, an instant transaction;

generating, with at least one processor, at least one feature value for the instant transaction;

determining, with at least one processor, that the at least one feature value corresponds to a feature associated with the account profile type; and

determining, with at least one processor, that the instant transaction is a non-suspect transaction based on the determination that the at least one feature value corresponds to a feature associated with the account profile type.

8. A system, comprising:

at least one processor programmed or configured to:

associate at least one transaction label from among a plurality of transaction labels with each of a plurality of recorded transactions, each transaction label associated with a transaction type;

generate a set of pre-processed feature values corresponding to a set of profile features based on the plurality of recorded transactions from a transaction profile and the transaction labels corresponding to each of the plurality of recorded transactions, the set of profile features having a first dimension;

process the set of pre-processed feature values with a macro-encoder to generate a first-reduced set of pre-processed feature values corresponding to a first-reduced set of profile features having a second dimension less than the first dimension;

cluster the first-reduced set of pre-processed feature values into at least two subsets of first-reduced feature values, a first subset of first-reduced feature values associated with a first macro-profile tag and a second subset of first-reduced feature values associated with a second macro-profile tag; and

tag the transaction profile with either the first macro-profile tag or the second macro-profile tag based on the clustering of the first-reduced set of feature values,

wherein the first macro-profile tag is associated with transaction profiles having features corresponding to low-threshold purchases, and the second macro-profile tag is associated with transaction profiles having features corresponding to high-threshold purchases.

9. The system according to claim 8 , wherein the at least one processor is further programmed or configured to:

determine an ordering of at least two subsets of first-reduced feature values.

10. The system according to claim 8 , wherein the at least one processor is further programmed or configured to:

process the set of pre-processed feature values with a micro-encoder to generate a second-reduced set of pre-processed feature values; and

cluster the second-reduced set of feature values into a plurality of subsets of second-reduced feature values, each subset of the second-reduced feature values associated with a micro-profile tag corresponding to a unique account profile type.

11. The system according to claim 9 , wherein the at least one processor is further programmed or configured to:

determine an ordering of the plurality of subsets of second-reduced feature values.

12. The system according to claim 8 , wherein each account profile type from among a plurality of account profile types is associated with features corresponding to a partially-unique set of behavior characteristics.

13. The system according to claim 12 , wherein the at least one processor is further programmed or configured to:

receive an instant transaction;

generate at least one feature value for the instant transaction;

determine that the at least one feature value does not correspond to a feature associated with the account profile type; and

determine that the instant transaction is a suspect transaction based on the determination that the at least one feature value does not correspond to a feature associated with the account profile type.

14. The system according to claim 12 , wherein the at least one processor is further programmed or configured to:

receive an instant transaction;

generate at least one feature value for the instant transaction;

determine that the at least one feature value corresponds to a feature associated with the account profile type; and

determine that the instant transaction is a non-suspect transaction based on the determination that the at least one feature value corresponds to a feature associated with the account profile type.

15. A computer program product, comprising at least one non-transitory computer-readable medium including program instructions that, when executed by at least one processor, cause the at least one processor to:

associate at least one transaction label from among a plurality of transaction labels with each of a plurality of recorded transactions, each transaction label associated with a transaction type;

generate a set of pre-processed feature values corresponding to a set of profile features based on the plurality of recorded transactions from a transaction profile and the transaction labels corresponding to each of the plurality of recorded transactions, the set of profile features having a first dimension;

process the set of pre-processed feature values with a macro-encoder to generate a first-reduced set of pre-processed feature values corresponding to a first-reduced set of profile features having a second dimension less than the first dimension;

cluster the first-reduced set of pre-processed feature values into at least two subsets of first-reduced feature values, a first subset of first-reduced feature values associated with a first macro-profile tag and a second subset of first-reduced feature values associated with a second macro-profile tag; and

tag the transaction profile with either the first macro-profile tag or the second macro-profile tag based on the clustering of the first-reduced set of feature values,

wherein the first macro-profile tag is associated with transaction profiles having features corresponding to low-threshold purchases, and the second macro-profile tag is associated with transaction profiles having features corresponding to high-threshold purchases.

16. The computer program product according to claim 15 , wherein the program instructions, when executed by the at least one processor, further cause the at least one processor to:

determine, with at least one processor, an ordering of at least two subsets of first-reduced feature values.

17. The computer program product according to claim 15 , wherein the program instructions, when executed by the at least one processor, further cause the at least one processor to:

process the set of pre-processed feature values with a micro-encoder to generate a second-reduced set of pre-processed feature values; and

cluster the second-reduced set of feature values into a plurality of subsets of second-reduced feature values, each subset of the second-reduced feature values associated with a micro-profile tag corresponding to a unique account profile type.

18. The computer program product according to claim 16 , wherein the program instructions, when executed by the at least one processor, further cause the at least one processor to:

determine an ordering of the plurality of subsets of second-reduced feature values.

19. The computer program product according to claim 15 , wherein each account profile type from among a plurality of account profile types is associated with features corresponding to a partially-unique set of behavior characteristics.

20. The computer program product according to claim 19 , wherein the program instructions, when executed by the at least one processor, further cause the at least one processor to:

receive an instant transaction;

generate at least one feature value for the instant transaction;

determine that the at least one feature value does not correspond to a feature associated with the account profile type; and

determine that the instant transaction is a suspect transaction based on the determination that the at least one feature value does not correspond to a feature associated with the account profile type.