Real-time predictive recommendation system using per-set optimization
In general, embodiments of the present invention provide systems, methods and computer readable media configured to use a per-set level optimization of the rank order of promotions to be recommended to a consumer. In some embodiments, machine learning is used offline to generate a predictive diversity model that receives one or more similarity rank features associated with a promotion (e.g., category, price band) as input, and produces an output multiplier to be applied to the promotion's respective associated relevance score (e.g., a relevance score representing a prediction of the promotion's conversion rate without diversity features). At run time, per-set optimization of the ordering of a set of promotions is implemented by adjusting the respective associated relevance scores of the promotions using the diversity model and then re-ordering the set of promotions based on their respective adjusted relevance scores.
1. A system, comprising one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to:
calculate a diversity multiplier for a promotion in an ordered set of promotions based at least in part on a mapping of a feature set associated with the ordered set of promotions and a probability class associated with the mapping of the feature set;
calculate a relevance score for the promotion based at least in part on the diversity multiplier and a consumer identifier associated with a consumer;
re-order the ordered set of promotions based at least in part on the relevance score for the promotion to generate a reordered set of promotions; and
transmit, to a consumer device associated with the consumer identifier, a top N ranked promotions from the reordered set of promotions based at least in part on respective relevance scores.
2. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:
calculate the diversity multiplier based at least in part on consumer behavior data associated with the consumer identifier.
3. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:
select the feature set associated with the ordered set of promotions based at least in part on one or more patterns with respect to features associated with the ordered set of promotions.
4. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:
select the feature set associated with the ordered set of promotions based at least in part on one or more promotion category relationships among features associated with the ordered set of promotions.
5. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:
select the feature set associated with the ordered set of promotions based at least in part on one or more location relationships among features associated with the ordered set of promotions.
6. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:
generate the mapping of the feature set based at least in part on historical consumer behavior data with respect to one or more promotions previously presented to one or more consumer devices.
7. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:
generate the mapping of the feature set based at least in part on historical promotion performance data with respect to one or more promotions previously presented to one or more consumer devices.
8. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:
calculate the diversity multiplier based at least in part on a neural network mapping of the feature set associated with the ordered set of promotions.
9. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:
calculate the diversity multiplier based at least in part on an ensemble of trees mapping of the feature set associated with the ordered set of promotions.
10. A computer-implemented method, comprising:
calculating, by a device comprising a processor, a diversity multiplier for a promotion in an ordered set of promotions based at least in part on a mapping of a feature set associated with the ordered set of promotions and a probability class associated with the mapping of the feature set;
calculating, by the device, a relevance score for the promotion based at least in part on the diversity multiplier and a consumer identifier associated with a consumer;
re-ordering, by the device, the ordered set of promotions based at least in part on the relevance score for the promotion to generate a reordered set of promotions; and
transmitting, by the device and to a consumer device associated with the consumer identifier, a top N ranked promotions from the reordered set of promotions based at least in part on respective relevance scores.
11. The computer-implemented method of claim 10 , wherein the calculating the diversity multiplier further comprises calculating the diversity multiplier based at least in part on consumer behavior data associated with the consumer identifier.
12. The computer-implemented method of claim 10 , wherein the computer-implemented method further comprises:
selecting, by the device, the feature set associated with the ordered set of promotions based at least in part on one or more patterns with respect to features associated with the ordered set of promotions.
13. The computer-implemented method of claim 10 , wherein the computer-implemented method further comprises:
selecting, by the device, the feature set associated with the ordered set of promotions based at least in part on one or more promotion category relationships among features associated with the ordered set of promotions.
14. The computer-implemented method of claim 10 , wherein the computer-implemented method further comprises:
selecting, by the device, the feature set associated with the ordered set of promotions based at least in part on one or more location relationships among features associated with the ordered set of promotions.
15. The computer-implemented method of claim 10 , wherein the computer-implemented method further comprises:
generating, by the device, the mapping of the feature set based at least in part on historical promotion performance data with respect to one or more promotions previously presented to one or more consumer devices.
16. A computer program product, stored on a computer readable medium, comprising instructions that when executed by one or more computers cause the one or more computers to:
calculate a diversity multiplier for a promotion in an ordered set of promotions based at least in part on a mapping of a feature set associated with the ordered set of promotions and a probability class associated with the mapping of the feature set;
calculate a relevance score for the promotion based at least in part on the diversity multiplier and a consumer identifier associated with a consumer;
re-order the ordered set of promotions based at least in part on the relevance score for the promotion to generate a reordered set of promotions; and
transmit, to a consumer device associated with the consumer identifier, a top N ranked promotions from the reordered set of promotions based at least in part on respective relevance scores.
17. The computer program product of claim 16 , wherein the computer program product further comprises instructions that when executed by the one or more computers cause the one or more computers to:
calculate the diversity multiplier based at least in part on consumer behavior data associated with the consumer identifier.