System, method, and computer program product for predicting consumer behavior based on demographics and new product features using machine learning models

1 . A system comprising:

at least one processor configured to:

receive spending behavior data, payment device feature data, and transaction data associated with each existing payment device of a plurality of existing payment devices, the spending behavior data associated with a plurality of spending behavior-related features of a respective group of users for each respective existing payment device of the plurality of existing payment devices, the payment device feature data associated with a plurality of payment device features for each existing payment device of the plurality of existing payment devices, the transaction data associated with at least one transaction related feature of the respective group of users for each respective existing payment device of the plurality of existing payment devices;

receive proposed spending behavior data and proposed payment device feature data associated with at least one proposed payment device, the proposed spending behavior data associated with a plurality of spending behavior-related features of a proposed group of users for the at least one proposed payment device, the proposed payment device feature data associated with a plurality of payment device features for the at least one proposed payment device;

train a plurality of machine learning models to predict a plurality of transaction metrics associated with the transaction data based on the spending behavior data and the payment device feature data associated with each existing payment device of the plurality of existing payment devices, wherein each respective machine learning model of the plurality of machine learning models is associated with a respective transaction metric of the plurality of transaction metrics, wherein, when training the plurality of machine learning models, the at least one processor is configured to:

calculate a loss for each of the plurality of machine learning models based on the plurality of transaction metrics and the transaction data;

update each of the plurality of machine learning models based on a respective loss by adjusting at least one trainable parameter of each of the plurality of machine learning models to provide a plurality of updated machine learning models;

determine whether each of the plurality of updated machine learning models satisfies a threshold level of accuracy; and

repeatedly update the plurality of updated machine learning models until the threshold level of accuracy is satisfied, wherein a plurality of trained machine learning models comprises the plurality of updated machine learning models upon determining that the plurality of updated machine learning models satisfies the threshold level of accuracy;

predict a plurality of predicted transaction metrics based on the proposed spending behavior data and the proposed payment device feature data associated with the at least one proposed payment device using the plurality of trained machine learning models;

determine a respective index for each of the plurality of transaction metrics from the plurality of trained machine learning models based on the plurality of predicted transaction metrics for the at least one proposed payment device and the plurality of transaction metrics predicted for the plurality of existing payment devices;

determine a plurality of final predicted transaction metrics based on the respective index for each of the plurality of transaction metrics from the plurality of trained machine learning models and a plurality of actual transaction metrics for the plurality of existing payment devices;

determine, based on an adjustment factor, a range for each final predicted transaction metric of the plurality of final predicted transaction metrics based on the respective index for each of the plurality of transaction metrics from the plurality of trained machine learning models;

determine a feature importance ranking of the plurality of payment device features;

generate at least one communication based on the range for each of the plurality of final predicted transaction metrics and the feature importance ranking; and

transmit the at least one communication generated based on the range for each of the plurality of final predicted transaction metrics and the feature importance ranking.

2 . The system of claim 1 , wherein the at least one processor is further configured to:

determine an average value of the at least one transaction related feature, wherein the at least one predicted transaction metric comprises a percentage change of the at least one transaction related feature relative to the average value of the at least one transaction related feature.

3 . The system of claim 1 , wherein the plurality of existing payment devices comprises a comparison payment device.

4 . The system of claim 3 , wherein the at least one predicted transaction metric is a percentage change of the at least one transaction related feature relative to an actual value of the at least one transaction related feature for the comparison payment device, and wherein the proposed spending behavior data comprises the spending behavior data associated with the comparison payment device.

5 . The system of claim 1 , wherein the spending behavior data comprises demographic data, issuer attribute data, and macroeconomic data.

6 . The system of claim 1 , wherein the at least one processor is further configured to:

match at least a subset of the plurality of payment device features of each existing payment device with at least a subset of the plurality of payment device features of the at least one proposed payment device,

wherein the plurality of existing payment devices are selected based on the matching.

7 . A method comprising:

receiving, with at least one processor, spending behavior data, payment device feature data, and transaction data associated with each existing payment device of a plurality of existing payment devices, the spending behavior data associated with a plurality of spending behavior-related features of a respective group of users for each respective existing payment device of the plurality of existing payment devices, the payment device feature data associated with a plurality of payment device features for each existing payment device of the plurality of existing payment devices, the transaction data associated with at least one transaction related feature of the respective group of users for each respective existing payment device of the plurality of existing payment devices;

receiving, with at least one processor, proposed spending behavior data and proposed payment device feature data associated with at least one proposed payment device, the proposed spending behavior data associated with a plurality of spending behavior-related features of a proposed group of users for the at least one proposed payment device, the proposed payment device feature data associated with a plurality of payment device features for the at least one proposed payment device;

training, with at least one processor, a plurality of machine learning models to predict a plurality of transaction metrics associated with the transaction data based on the spending behavior data and the payment device feature data associated with each existing payment device of the plurality of existing payment devices, wherein each respective machine learning model of the plurality of machine learning models is associated with a respective transaction metric of the plurality of transaction metrics, wherein training the plurality of machine learning models comprises:

calculating a loss for each of the plurality of machine learning models based on the plurality of transaction metrics and the transaction data;

updating each of the plurality of machine learning models based on a respective loss by adjusting at least one trainable parameter of each of the plurality of machine learning models to provide a plurality of updated machine learning models;

determining whether each of the plurality of updated machine learning models satisfies a threshold level of accuracy; and

repeatedly updating the plurality of updated machine learning models until the threshold level of accuracy is satisfied, wherein a plurality of trained machine learning models comprises the plurality of updated machine learning models upon determining that the plurality of updated machine learning models satisfies the threshold level of accuracy;

predicting, with at least one processor, a plurality of predicted transaction metrics based on the proposed spending behavior data and the proposed payment device feature data associated with the at least one proposed payment device using the plurality of trained machine learning models;

determining, with at least one processor, a respective index for each of the plurality of transaction metrics from the plurality of trained machine learning models based on the plurality of predicted transaction metrics for the at least one proposed payment device and the plurality of transaction metrics predicted for the plurality of existing payment devices;

determining, with at least one processor, a plurality of final predicted transaction metrics based on the respective index for each of the plurality of transaction metrics from the plurality of trained machine learning models and a plurality of actual transaction metrics for the plurality of existing payment devices;

determining, with at least one processor, based on an adjustment factor, a range for each final predicted transaction metric of the plurality of final predicted transaction metrics based on the respective index for each of the plurality of transaction metrics from the plurality of trained machine learning models;

determining, with at least one processor, a feature importance ranking of the plurality of payment device features;

generate at least one communication based on the range for each of the plurality of final predicted transaction metrics and the feature importance ranking; and

transmitting, with at least one processor, the at least one communication generated based on the range for each of the plurality of final predicted transaction metrics and the feature importance ranking.

8 . The method of claim 7 , further comprising:

determining, with at least one processor, an average value of the at least one transaction related feature, wherein the at least one predicted transaction metric comprises a percentage change of the at least one transaction related feature relative to the average value of the at least one transaction related feature.

9 . The method of claim 7 , wherein the plurality of existing payment devices comprises a comparison payment device, wherein the at least one predicted transaction metric is a percentage change of the at least one transaction related feature relative to an actual value of the at least one transaction related feature for the comparison payment device, and wherein the proposed spending behavior data comprises the spending behavior data associated with the comparison payment device.

10 . The method of claim 7 , wherein the spending behavior data comprises demographic data, issuer attribute data, and macroeconomic data.

11 . The method of claim 7 , further comprising:

matching, with at least one processor, at least a subset of the plurality of payment device features of each existing payment device with at least a subset of the plurality of payment device features of the at least one proposed payment device,

wherein the plurality of existing payment devices are selected based on the matching.

12 . 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:

receive spending behavior data, payment device feature data, and transaction data associated with each existing payment device of a plurality of existing payment devices, the spending behavior data associated with a plurality of spending behavior-related features of a respective group of users for each respective existing payment device of the plurality of existing payment devices, the payment device feature data associated with a plurality of payment device features for each existing payment device of the plurality of existing payment devices, the transaction data associated with at least one transaction related feature of the respective group of users for each respective existing payment device of the plurality of existing payment devices;

receive proposed spending behavior data and proposed payment device feature data associated with at least one proposed payment device, the proposed spending behavior data associated with a plurality of spending behavior-related features of a proposed group of users for the at least one proposed payment device, the proposed payment device feature data associated with a plurality of payment device features for the at least one proposed payment device;

train a plurality of machine learning models to predict a plurality of transaction metrics associated with the transaction data based on the spending behavior data and the payment device feature data associated with each existing payment device of the plurality of existing payment devices, wherein each respective machine learning model of the plurality of machine learning models is associated with a respective transaction metric of the plurality of transaction metrics wherein, the instructions that cause the at least one processor to train the plurality of machine learning models, cause the at least one processor to:

calculate a loss for each of the plurality of machine learning models based on the plurality of transaction metrics and the transaction data;

update each of the plurality of machine learning models based on a respective loss by adjusting at least one trainable parameter of each of the plurality of machine learning models to provide a plurality of updated machine learning models;

determine whether each of the plurality of updated machine learning models satisfies a threshold level of accuracy; and

repeatedly update the plurality of updated machine learning models until the threshold level of accuracy is satisfied, wherein a plurality of trained machine learning models comprises the plurality of updated machine learning models upon determining that the plurality of updated machine learning models satisfies the threshold level of accuracy;

predict a plurality of predicted transaction metrics based on the proposed spending behavior data and the proposed payment device feature data associated with the at least one proposed payment device using the plurality of trained machine learning models;

determine a respective index for each of the plurality of transaction metrics from the plurality of trained machine learning models based on the plurality of predicted transaction metrics for the at least one proposed payment device and the plurality of transaction metrics predicted for the plurality of existing payment devices;

determine a plurality of final predicted transaction metrics based on the respective index for each of the plurality of transaction metrics from the plurality of trained machine learning models and a plurality of actual transaction metrics for the plurality of existing payment devices;

determine, based on an adjustment factor, a range for each final predicted transaction metric of the plurality of final predicted transaction metrics based on the respective index for each of the plurality of transaction metrics from the plurality of trained machine learning models;

determine a feature importance ranking of the plurality of payment device features;

generate at least one communication based on the range for each of the plurality of final predicted transaction metrics and the feature importance ranking; and

transmit the at least one communication generated based on the range for each of the plurality of final predicted transaction metrics and the feature importance ranking.

13 . The computer program product of claim 12 , wherein the program instructions, when executed by the at least one processor, further cause the at least one processor to:

determine an average value of the at least one transaction related feature, wherein the at least one predicted transaction metric comprises a percentage change of the at least one transaction related feature relative to the average value of the at least one transaction related feature.

14 . The computer program product of claim 12 , wherein the plurality of existing payment devices comprises a comparison payment device, wherein the at least one predicted transaction metric is a percentage change of the at least one transaction related feature relative to an actual value of the at least one transaction related feature for the comparison payment device, and wherein the proposed spending behavior data comprises the spending behavior data associated with the comparison payment device.

15 . The computer program product of claim 12 , wherein the program instructions, when executed by the at least one processor, further cause the at least one processor to:

match at least a subset of the plurality of payment device features of each existing payment device with at least a subset of the plurality of payment device features of the at least one proposed payment device,

wherein the plurality of existing payment devices are selected based on the matching.