Systems and methods for intelligently optimizing a queue of actions in an interface using machine learning

A system, comprising:

• one or more databases including training data stored therein;
• one or more memory devices storing instructions;
• one or more processors configured to execute the instructions;
• and
• one or more interfaces for managing interactions between the one or more processors and the one or more databases,
• the one or more processors configured to execute the instructions to perform operations comprising:
• analyzing the training data stored in the one or more databases to build a predictive model;
• applying the predictive model to a plurality of loan applications to determine for each loan application of the plurality of loan applications:
• a first probability of an institution funding the loan application if the institution proactively engages a loan-arranging entity of the loan application;
• and
• a second probability of the institution funding the loan application if the institution reactively engages the loan-arranging entity;
• reporting the first and second probabilities to a user through a user interface of the one or more interfaces;
• updating the predictive model;
• determining a value of each loan application of the plurality of loan applications by:
• determining differences between the first and second probabilities for each respective loan application;
• and
• multiplying the differences by scalars provided for the loan applications to determine scaled values;
• ranking the plurality of loan applications based on the first and second probabilities;
• and
• reporting the plurality of loan applications in a ranking order to the user through the user interface,
• wherein the analyzing the training data includes building the predictive model utilizing a machine-learning system trained using the training data, and
• wherein the machine-learning system comprises a gradient boosting machine.

A method, comprising:

• analyzing training data stored in a database to build a predictive model;
• applying the predictive model to a plurality of loan applications to determine for each loan application of the plurality of loan applications:
• a first probability of an institution funding the loan applications if the institution proactively engages loan-arranging entities of the loan applications;
• and
• a second probability of the institution funding the loan applications if the institution reactively engages the loan-arranging entities of the loan applications;
• updating the predictive model;
• ranking the loan applications based on the first and second probabilities;
• reporting the loan applications in a ranking order to a user through a user interface;
• and
• determining a value of each loan application of the plurality of loan applications by:
• determining differences between the first and second probabilities for each respective loan application;
• and
• multiplying the differences by scalars provided for the loan applications to determine scaled values,
• wherein the analyzing the training data includes building the predictive model utilizing a machine-learning system trained using the training data, and
• wherein the machine-learning system comprises a gradient boosting machine.

A non-transitory memory storing instructions that, when executed by at least one processor, cause a system to perform operations comprising:

• analyzing training data stored in a database to build a predictive model;
• applying the predictive model to a plurality of loan applications, to determine for each loan application of the plurality of loan applications:
• a first probability of an institution funding the loan applications if the institution proactively engages loan-arranging entities of the loan applications;
• and
• a second probability of the institution funding the loan applications if the institution reactively engages the loan-arranging entities of the loan applications;
• updating the predictive model;
• ranking the loan applications based on the first and second probabilities;
• reporting the loan applications in a ranking order to a user through a user interface;
• determining a value of each loan application of the plurality of loan applications by:
• determining differences between the first and second probabilities for each respective loan application;
• and
• multiplying the differences by scalars provided for the loan applications to determine scaled values;
• and
• reporting the respective values to the user through the user interface;
• wherein the analyzing the training data includes building the predictive model utilizing a machine-learning system trained using the training data, and
• wherein the machine-learning system comprises a gradient boosting machine.