Adaptive recommendation system and methods

An adaptive recommendation (“AR”) computer device used for predicting future purchases of a consumer based on consumer action, said AR computer device comprising a memory device and a processor communicatively coupled to the memory device, said processor programmed to:

• receive at least one preference vector from a recommender server computing device, wherein the at least one preference vector represents a purchasing pattern of a plurality of cardholders based on a first plurality of payment transactions corresponding to at least one item category, the first plurality of payment transactions from the plurality of cardholders;
• store, in the memory device, the at least one preference vector;
• receive, for a candidate consumer, a plurality of consumer actions each including a time aspect, the plurality of consumer actions comprising a second plurality of payment transactions and a plurality of online actions, the second plurality of payment transaction from the candidate consumer, wherein the time aspect includes at least a time when each consumer action of the plurality of consumer actions occurred;
• alter, using at least one adaptive learning algorithm, the at least one preference vector based on the plurality of consumer actions;
• determine at least one personalized vector for the candidate consumer based on the altered at least one preference vector;
• adjust, using the at least one adaptive learning algorithm, the at least one personalized vector such that more recent consumer actions of the plurality of consumer actions have a greater impact than less recent consumer actions of the plurality of consumer actions;
• store, in the memory device, the adjusted at least one personalized vector;
• synchronize, with at least one other AR computer device, the adjusted at least one personalized vector by transmitting the adjusted at least one personalized vector to the at least one other AR computer device such that the adjusted at least one personalized vector is stored within the at least one other AR computer device;
• receive a deletion signal from a remote computer device associated with the candidate consumer;
• delete, from the memory device in response to the deletion signal, the at least one personalized vector;
• restore, from the at least one other AR computer device, the adjusted at least one personalized vector, wherein the adjusted at least one personalized vector has been further altered by actions performed on the at least one other AR computer device;
• receive a plurality of local filters, the plurality of local filters including a current time, a current geographic information of the candidate consumer, and one or more recent consumer actions associated with the candidate consumer based on the current time;
• apply the plurality of local filters to the at least one personalized vector;
• generate at least one recommendation for a future purchase in the at least one item category based on the plurality of local filters applied to the at least one personalized vector;
• and
• display the at least one recommendation to the candidate consumer.

At least one non-transitory computer-readable storage media having computer-executable instructions embodied thereon, wherein when executed by an adaptive recommendation computer device having at least one memory device and at least one processor coupled to the at least one memory device, the computer-executable instructions cause the processor to:

• receive at least one preference vector from a recommender server computing device, wherein the at least one preference vector represents a purchasing pattern of a plurality of cardholders based on a first plurality of payment transactions corresponding to at least one item category, the first plurality of payment transactions from the plurality of cardholders;
• store, in the at least one memory device, the at least one preference vector;
• receive, for a candidate consumer, a plurality of consumer actions each including a time aspect, the plurality of consumer actions comprising a second plurality of payment transactions and a plurality of online actions, the second plurality of payment transaction from the candidate consumer, wherein the time aspect includes at least a time when each consumer action of the plurality of consumer actions occurred;
• alter, using at least one adaptive learning algorithm, the at least one preference vector based on the plurality of consumer actions;
• determine at least one personalized vector for the candidate consumer based on the altered at least one preference vector;
• adjust, using the at least one adaptive learning algorithm, the at least one personalized vector such that more recent consumer actions of the plurality of consumer actions have a greater impact than less recent consumer actions of the plurality of consumer actions;
• store, in the at least one memory device, the adjusted at least one personalized vector;
• synchronize, with at least one other AR computer device, the adjusted at least one personalized vector by transmitting the adjusted at least one personalized vector to the at least one other AR computer device such that the adjusted at least one personalized vector is stored within the at least one other AR computer device;
• receive a deletion signal from a remote computer device associated with the candidate consumer;
• delete, from the at least one memory device in response to the deletion signal, the at least one personalized vector;
• restore, from the at least one other AR computer device, the adjusted at least one personalized vector, wherein the adjusted at least one personalized vector has been further altered by actions performed on the at least one other AR computer device;
• receive a plurality of local filters, the plurality of local filters including a current time, a current geographic information of the candidate consumer, and one or more recent consumer actions associated with the candidate consumer based on the current time;
• apply the plurality of local filters to the at least one personalized vector;
• generate at least one recommendation for a future purchase in the at least one item category based on the plurality of local filters applied to the at least one personalized vector;
• and
• display the at least one recommendation to the candidate consumer.