Automated data forecasting using machine learning

1. A system comprising:

one or more processors; and

a memory having programming instructions stored thereon, which, when executed by the one or more processors, performs operations, comprising:

retrieving historical account activity for a plurality of users, the historical account activity comprising historical inflow data and historical outflow data;

constructing a training data set comprising the historical inflow data, the historical outflow data, and known forecast information from the historical account activity;

generating a prediction model configured to forecast future inflow activity and future outflow activity, the generating comprising:

learning, by the prediction model, to forecast the future inflow activity and the future outflow activity based on the training data set, the learning comprising optimizing an objective function of the prediction model by penalizing errors for projected inflow predictions, projected outflow predictions, and differences between the projected inflow predictions and projected outflow predictions to reduce a drift in forecasted values,

the objective function comprising at least one weight parameter that, if adjusted in favor of a predicted account balance, increases an accuracy of the predicted account balance and decreases an accuracy of the projected inflow predictions and the projected outflow predictions;

receiving, from one or more third-party systems, current inflow activity, current outflow activity, and current balance information for a user;

generating the predicted account balance by:

forecasting, by the prediction model, a future inflow and a future outflow; and

constructing the predicted account balance based on the future inflow, the future outflow, and the current balance information;

retraining the prediction model, said retraining comprising:

evaluating a performance of the prediction model based on an accuracy of one or more of the future inflow, the future outflow and the predicted account balance,

adjusting the at least one weight parameter based on said evaluating, and

optimizing the objective function based on the adjusted at least one weight parameter.

2. The system of claim 1 , wherein the objective function comprises:

an inflow component associated with the projected inflow predictions;

an outflow component associated with the projected outflow predictions; and

a difference component associated with the differences between the projected inflow predictions and the projected outflow predictions.

3. The system of claim 2 , wherein the weight parameter provides a tradeoff between the projected inflow predictions, the projected outflow predictions, and the difference component.

4. The system of claim 1 , wherein adjusting the at least one weight parameter in favor of the projected inflow predictions and the projected outflow predictions increases an accuracy of the projected inflow predictions and the projected outflow predictions and decreases an accuracy of the predicted account balance.

5. The system of claim 1 , wherein the operations further comprise:

generating a graphical user interface comprising a dashboard capturing account information associated with the user, the account information comprising the future inflow, the future outflow, and the predicted account balance.

6. The system of claim 1 , wherein constructing the training data set comprises:

deriving transaction type information based on the historical account activity; and

augmenting the training data set with the transaction type information.

7. The system of claim 1 , wherein the operations further comprise:

generating a plurality of insights based on at least one of the future inflow, the future outflow, and the predicted account balance.

8. The system of claim 1 ,

wherein retraining the prediction model occurs at least one of periodically and continuously, and wherein the evaluating occurs automatically.

9. A method comprising:

retrieving, by a computing system, historical account activity for a plurality of users, the historical account activity comprising historical inflow data and historical outflow data, the computing system comprising one or more processors executing programming instructions stored in a memory;

constructing, by the computing system, a training data set comprising the historical inflow data, the historical outflow data, and known forecast information from the historical account activity; and

generating, by the computing system, a prediction model configured to forecast future inflow activity and future outflow activity, the generating comprising:

learning, by the prediction model, to forecast the future inflow activity and the future outflow activity based on the training data set, the learning comprising optimizing an objective function of the prediction model by penalizing errors for projected inflow predictions, projected outflow predictions, and differences between the projected inflow predictions and projected outflow predictions to reduce a drift in forecasted values,

the objective function comprising at least one weight parameter that, if adjusted in favor of the differences, increases an accuracy of the differences and decreases an accuracy of the projected inflow predictions and the projected outflow predictions;

retraining, by the computing system, the prediction model, said retraining comprising:

evaluating a performance of the prediction model based on an accuracy of one or more of a future inflow, a future outflow and a predicted account balance generated by the prediction model,

adjusting the at least one weight parameter based on said evaluating, and

optimizing the objective function based on the adjusted at least one weight parameter.

10. The method of claim 9 , wherein the objective function comprises:

an inflow component associated with the projected inflow predictions;

an outflow component associated with the projected outflow predictions; and

a difference component associated with the differences between the projected inflow predictions and the projected outflow predictions.

11. The method of claim 10 , wherein the weight parameter provides a tradeoff between the projected inflow predictions, the projected outflow predictions, and the differences.

12. The method of claim 9 , wherein adjusting the at least one weight parameter in favor of the projected inflow predictions and the projected outflow predictions increases an accuracy of the projected inflow predictions and the projected outflow predictions and decreases an accuracy of the differences.

13. The method of claim 9 , wherein constructing the training data set comprises:

deriving transaction type information based on the historical account activity; and

augmenting the training data set with the transaction type information.

14. The method of claim 9 , wherein constructing the training data set comprises:

accessing a userbase comprising the plurality of users; and

clustering the userbase into a plurality of clusters based on demographic information associated with the plurality of users or characteristic information associated with a plurality of organizations.

15. The method of claim 9 ,

wherein retraining the prediction model occurs at least one of periodically and continuously, and wherein the evaluating occurs automatically.

16. The method of claim 9 , wherein constructing the training data set comprises:

augmenting the training data set with exogenous inputs comprising one or more of calendar data related to the historical inflow data, the historical outflow data and customer characteristic information.

17. A non-transitory computer readable medium comprising one or more sequences of instructions, which, when executed by one or more processors, causes a computing system to perform operations comprising:

receiving, by the computing system, current inflow activity, current outflow activity, and current balance information for a user;

generating, by the computing system, future forecasts for the user using a prediction model trained to forecast future inflow activity and forecast future outflow activity by:

forecasting, by the prediction model, a future inflow and a future outflow based on the current inflow activity, current outflow activity, and the current balance information for the user;

constructing, by the computing system, future balance data based on the future inflow and the future outflow; and

presenting, by the computing system, the future inflow, the future outflow, and the future balance data to the user,

wherein the prediction model is trained by optimizing an objective function that includes at least one weight parameter, and wherein adjusting the at least one weight parameter in favor of the future balance data increases an accuracy of the future balance data and decreases an accuracy of the future inflow and the future outflow; and

retraining the prediction model, said retraining comprising:

evaluating a performance of the prediction model based on an accuracy of one or more of the future inflow, the future outflow and the future balance data;

adjusting the at least one weight parameter based on said evaluating and

optimizing the objective function based on the adjusted at least one weight parameter.

18. The non-transitory computer readable medium of claim 17 , further comprising:

generating a graphical user interface comprising a dashboard capturing account information associated with the user, the account information comprising the future inflow, the future outflow, and the future balance data.