Optimizing neural networks for risk assessment

Patent metadata

Field	Value
Publication number	US-10133980-B2
Application number	US-201615560401-A
Country	US
Kind code	B2
Filing date	Mar 25, 2016
Priority date	Mar 27, 2015
Publication date	Nov 20, 2018
Grant date	Nov 20, 2018

Abstract

Certain embodiments involve generating or optimizing a neural network for risk assessment. The neural network can be generated using a relationship between various predictor variables and an outcome (e.g., a condition's presence or absence). The neural network can be used to determine a relationship between each of the predictor variables and a risk indicator. The neural network can be optimized by iteratively adjusting the neural network such that a monotonic relationship exists between each of the predictor variables and the risk indicator. The optimized neural network can be used both for accurately determining risk indicators using predictor variables and determining adverse action codes for the predictor variables, which indicate an effect or an amount of impact that a given predictor variable has on the risk indicator. The neural network can be used to generate adverse action codes upon which consumer behavior can be modified to improve the risk indicator score.

First claim

What is claimed is:

1. A system comprising:
a network interface configured for communicating with a user device;
a processing device; and
a memory device in which instructions executable by the processing device are stored for causing the processing device to:
configure a neural network to generate risk indicators and adverse action codes, the adverse action codes indicating impacts of respective predictor variables on the risk indicators, wherein the configuration of the neural network includes operations comprising:
(a) retrieving, from a database in the system, a plurality of predictor variables, wherein each predictor variable corresponds to an entity,
(b) determining a correlation between each predictor variable and an outcome,
(c) generating the neural network, the neural network having a hidden layer for determining a relationship between each predictor variable and a risk indicator based on the correlation, wherein the risk indicator is a level of risk associated with the entity and wherein the neural network is operable for determining whether a monotonic relationship exists between each predictor variable and the risk indicator, and
(d) iteratively adjusting the neural network so that the monotonic relationship exists between each predictor variable and the risk indicator as determined by the neural network,
wherein each adjustment comprises adjusting at least one of a number of nodes in the hidden layer of the neural network, a predictor variable in the plurality of predictor variables, or a number of layers in the neural network,
wherein the processing device is configured to determine, based on a rate of change of the risk indicator with respect to each predictor variable, that the monotonic relationship exists between the predictor variable and the risk indicator, and

generate and provide, after the monotonic relationship exists between each predictor variable and the risk indicator, an output risk indicator and output set of adverse action codes by performing additional operations comprising:
(a) retrieving, from the database, an input set of predictor values for the predictor variables, wherein the input set of predictor values corresponds to a target entity,
(b) determining the output risk indicator by applying the iteratively adjusted neural network to the input set of predictor values,
(c) determining, using the neural network, impacts of the predictor variables, respectively, on the output risk indicator,
(d) generating, using the neural network, the output set of adverse action codes that respectively indicate the impacts of the predictor variables on the output risk indicator,
(e) generating an electronic communication that includes the output risk indicator and the output set of adverse action codes, and
(f) configuring the network interface to transmit the electronic communication to the user device for display of the output risk indicator and the output set of adverse action codes at the user device.

2. The system of claim 1 , wherein the hidden layer comprises at least two hidden layers.

3. The system of claim 1 , wherein the processing device is configured to determine the correlation between each predictor variable and an outcome by determining a correlation between each predictor variable and an amount of positive outcomes or negative outcomes, wherein each positive outcome indicates that a condition is satisfied and each negative outcome indicates failure to satisfy the condition.

4. The system of claim 3 , wherein the processing device is configured to determine the correlation between each predictor variable and an amount of positive outcomes or negative outcomes by performing operations comprising verifying that a bivariate relationship exists between each predictor variable and the amount of positive or negative outcomes.

5. The system of claim 1 , wherein the processing device is further configured to determine ranks of the predictor variables, respectively, using the neural network, based on the impacts of the predictor variables on the output risk indicator.

6. The system of claim 1 , wherein the risk indicator corresponds to a credit score of the entity.

7. A method that includes one or more processing devices performing operations comprising:
configuring a neural network to generate risk indicators and adverse action codes, the adverse action codes indicating impacts of respective predictor variables on the risk indicators, wherein the configuring of the neural network comprises:
(a) retrieving, from a database, a plurality of predictor variables, wherein each predictor variable corresponds to an entity,
(b) determining a correlation between each predictor variable and an amount of positive outcomes or negative outcomes, wherein each positive outcome indicates that a condition is satisfied and each negative outcome indicates failure to satisfy the condition,
(c) generating the neural network, the neural network having a hidden layer for determining a relationship between each predictor variable and a risk indicator based on the correlation, wherein the risk indicator is a level of risk associated with the entity, and
(d) iteratively adjusting the neural network so that a monotonic relationship exists between each predictor variable and the risk indicator as determined by the neural network,
wherein each adjustment comprises adjusting at least one of a number of nodes in the hidden layer of the neural network, a predictor variable in the plurality of predictor variables, or a number of layers in the neural network,
wherein the one or more processing devices determine, based on a rate of change of the risk indicator with respect to each predictor variable, that the monotonic relationship exists between the predictor variable and the risk indicator; and
generating and providing, after the monotonic relationship exists between each predictor variable and the risk indicator, an output risk indicator and output set of adverse action codes, wherein the generating and providing comprises:
(a) retrieving, from the database, an input set of predictor values for the predictor variables, wherein the input set of predictor values corresponds to a target entity,
(b) determining the output risk indicator by applying the iteratively adjusted neural network to the input set of predictor values,
(c) determining, using the neural network, impacts of the predictor variables, respectively, on the output risk indicator,
(d) generating, using the neural network, the output set of adverse action codes that respectively indicate the impacts of the predictor variables on the output risk indicator,
(e) generating an electronic communication that includes the output risk indicator and the output set of adverse action codes, and
(f) configuring a network interface to transmit the electronic communication from the one or more processing devices to a user device for display of the output risk indicator and the output set of adverse action codes at the user device.

8. The method of claim 7 , the operations further comprising determining, with the iteratively adjusted neural network, ranks for the predictor variables, respectively and outputting the ranks of the predictor variables.

9. The method of claim 7 , wherein the hidden layer comprises at least two hidden layers.

10. The method of claim 7 , wherein determining the correlation between each predictor variable and an amount of positive outcomes or negative outcomes includes verifying that a bivariate relationship exists between each predictor variable and the amount of positive or negative outcomes.

11. A non-transitory computer-readable storage medium having program code that is executable by a process

Assignees

• Equifax Inc
• Equifax Inc

Inventors

• Turner Matthew
• Mcburnett Michael

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06Q40/03

  Credit; Loans; Processing thereof · CPC title

• G06F17/18

  for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)} · CPC title

• G06Q40/00

  Finance; Insurance; Tax strategies; Processing of corporate or income taxes · CPC title

• G06Q40/08

  Insurance · CPC title