Systems and methods for expediting rule-based data processing

• US 10,599,985 B2
• Filed: 09/01/2017
• Issued: 03/24/2020
• Est. Priority Date: 09/01/2017
• Status: Active Grant

• Current Assignee: Capital One Services LLC (Capital One Financial Corporation)
• Original Assignee: Capital One Services LLC (Capital One Financial Corporation)
• Inventors: Crable, Fredrick A., El-Ali, Hala S.
• Primary Examiner(s): Vincent, David R
• Application Number:
  US15/693,994
• Publication Number:
  US 20190073599A1
• Time in Patent Office: 935 Days
• Field of Search: 706 15, 706 45, 706 47
• CPC Class Codes: G06F 16/9027 Trees
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/046 Forward inferencing; Production systems
• CPC Class Codes: G06Q 40/03 Credit; Loans; Processing thereof

Abstract

A system includes one or more memory devices storing instructions, and one or more processors configured to execute the instructions to perform steps of a method for generating a data tree structure for expediting rule-based decision-making. The system may receive a rules file having a plurality of rules, where each rule has one or more conditions and one or more actions. The system may generate a plurality of condition frequency counts that represent the total number of rules having a condition of a particular unique condition type. The system may then generate a tree structure based on the plurality of rules and the plurality of condition frequency counts, where the tree structure includes a root node and one or more node branches, where each node branch includes a first-level node and zero or more second-level nodes that each include a node condition and zero or more node actions.

Claims

1. A system for generating a data tree structure for expediting rule-based decision-making, the system comprising:

one or more processors; and

a memory in communication with the one or more processors and storing instructions that, when executed by the one or more processors, are configured to cause the system to;

receive a rules file comprising a plurality of rules, each rule of the plurality of rules comprising one or more conditions and one or more actions, wherein each condition is associated with a condition type;

responsive to reading the rules file, generate a plurality of condition frequency counts, each of the plurality of condition frequency counts being associated with a unique condition type, and each condition frequency count representing the total number of rules having a condition associated with the respective unique condition type;

generate, based on the plurality of rules and the plurality of condition frequency counts, the data tree structure comprising a root node and one or more node branches, each node branch comprising a first-level node and one or more second-level nodes, each of the first-level nodes and the one or more second-level nodes having a node condition and at least some of the first-level nodes and second-level nodes having one or more node actions;

receive an input request comprising one or more input conditions;

based on the one or more input conditions, traverse each first-level node of the one or more node branches;

determine a subset of the first-level nodes having conditions that are satisfied by one or more input conditions of the input request;

traverse to each second-level node of the one or more node branches that is linked to a node of the subset of first-level nodes, bypassing each second-level node of the one or more node branches that is not linked to a node of the subset of first-level nodes; and

execute one or more node actions associated with at least some of the traversed nodes.

2. The system of claim 1, the memory storing further instructions that, when executed by the one or more processors, are further configured to cause the system to:

identify a sub-group of traversed nodes, wherein each traversed node of the sub-group has a condition that is met by the one or more input conditions of the input request; and

execute the one or more node actions associated with at least some traversed nodes of the sub-group of traversed nodes.

3. The system of claim 1, wherein at least one first-level node is formed from one rule.

4. The system of claim 1, wherein at least one first-level node is formed from two rules.

5. The system of claim 1, wherein the data tree structure is a hierarchy of singly linked nodes.

6. The system of claim 2, wherein one or more of the node actions have a respective associated priority, and wherein the instructions, when executed by the one or more processors, are further configured to cause the system to, responsive to determining that two or more of the node actions associated with the sub-group conflict with one another, prioritize execution of the two or more conflicting node actions based on their respective associated priorities.

7. The system of claim 1, wherein the condition types comprise one or more of line-of-business in, tier in, and version.

8. The system of claim 1, wherein generating the data tree structure based on the plurality of rules and the plurality of condition frequency counts comprises:

generating a root node;

prioritizing the condition types based on the magnitude of each condition type'"'"'s respective frequency count;

processing a first first-rule condition of a first rule by generating a first child node to the root node, the first first-rule condition having a condition type of the highest priority and the first child node having the first first-rule condition;

processing one or more first-rule conditions of the first rule in order of their condition type priority by, for each first-rule condition, creating a first-rule child node including the first-rule condition, where the first-rule child node is a child node of the previously created first-rule node, and for the last processed condition, incorporating one or more actions into the last created first-rule child node; and

processing a second rule by;

responsive to determining that a first second-rule condition is not included in the first child node to the root node, generating a second child node to the root node incorporating the first second-rule condition, wherein the first second-rule condition has a condition type of the highest priority;

responsive to determining that the first second-rule condition is included in the first child node to the root node and that a child node of the first child node to the root node does not include a second second-rule condition, creating a second-rule child node including the second-rule condition, the second-rule child node being a child of the first child node to the root node.

9. A computer-implemented method for generating a data tree structure to be used for expediting rule-based decision-making, the method comprising:

receiving a rules file comprising a plurality of rules, each rule of the plurality of rules comprising one or more conditions and one or more actions, wherein each condition is associated with a condition type;

responsive to reading the rules file, generating a plurality of condition frequency counts, each of the plurality of condition frequency counts being associated with a unique condition type, and each condition frequency count representing the total number of rules having a condition associated with the respective unique condition type;

generating, based on the plurality of rules and the plurality of condition frequency counts, the data tree structure comprising a root node and one or more node branches, each node branch comprising a first-level node and one or more second-level nodes, each of the first-level nodes and the one or more second-level nodes having a node condition and at least some of the first-level nodes and second-level nodes having one or more node actions;

receiving an input request comprising one or more input conditions;

based on the input conditions, traverse each first-level node of the one or more node branches;

determine a subset of the first-level nodes having conditions that are satisfied by one or more input conditions of the input request;

traverse to each second-level node of the one or more node branches that is linked to a node of the subset of first-level nodes, bypassing each second-level node of the one or more node branches that is not linked to a node of the subset of first-level nodes; and

execute one or more node actions associated with at least some of the traversed nodes.

10. The method of claim 9, further comprising:

identifying a sub-group of traversed nodes, wherein each traversed node of the sub-group includes a condition that is met by the one or more input conditions of the input request; and

executing the one or more node actions associated with at least some traversed nodes of the sub-group of traversed nodes.

11. The method of claim 9, wherein at least one first-level node is formed from one rule.

12. The method of claim 9, wherein at least one first-level node is formed from two rules.

13. A non-transitory computer-readable medium storing instructions for generating a data tree structure to be used for expediting rule-based decision-making that, when executed by one or more processors, cause a system to:

receive a rules file comprising a plurality of rules, each rule of the plurality of rules comprising one or more conditions and one or more actions, wherein each condition is associated with a condition type;

responsive to reading the rules file, generate a plurality of condition frequency counts, each of the plurality of condition frequency counts being associated with a unique condition type, and each condition frequency count representing the total number of rules having a condition associated with the respective unique condition type;

generate, based on the plurality of rules and the plurality of condition frequency counts, the data tree structure comprising a root node and one or more node branches, each node branch comprising a first-level node and one or more second-level nodes, each of the first-level nodes and the one or more second-level nodes having a node condition and at least some of the first-level nodes and second-level nodes having one or more node actions;

receive an input request comprising one or more input conditions;

based on the one or more input conditions, traverse each first-level node of the one or more node branches;

determine a subset of the first-level nodes having conditions that are satisfied by one or more input conditions of the input request;

traverse to each second-level node of the one or more node branches that is linked to a node of the subset of first-level nodes, bypassing each second-level node of the one or more node branches that is not linked to a node of the subset of first-level nodes; and

execute one or more node actions associated with at least some of the traversed nodes.

14. The non-transitory computer-readable medium of claim 13, wherein the data tree structure is a hierarchy of singly linked nodes.

15. The non-transitory computer-readable medium of claim 13, wherein condition types comprise one or more of line-of-business in, tier in, and version.