US 11,810,001 B1
Systems and methods for generating and implementing knowledge graphs for knowledge representation and analysis
Yuang Tang, Baltimore, MD (US); Fabio Quijada, Reston, VA (US); and Dylan Nielson, Washington, DC (US)
Assigned to FEDERAL HOME LOAN MORTGAGE CORPORATION (FREDDIE MAC), McLean, VA (US)
Filed by Federal Home Loan Mortgage Corporation (Freddie Mac), McLean, VA (US)
Filed on Dec. 18, 2018, as Appl. No. 16/224,786.
Application 16/224,786 is a continuation in part of application No. 15/593,113, filed on May 11, 2017, granted, now 10,496,678.
Claims priority of provisional application 62/335,580, filed on May 12, 2016.
Int. Cl. G06N 5/02 (2023.01); G06F 16/9032 (2019.01); G06F 17/18 (2006.01); G06N 20/00 (2019.01); G06F 18/231 (2023.01)
CPC G06N 5/02 (2013.01) [G06F 16/9032 (2019.01); G06F 17/18 (2013.01); G06F 18/231 (2023.01); G06N 20/00 (2019.01)] 7 Claims
OG exemplary drawing
 
1. A knowledge graph computer system, comprising:
at least one processor;
at least one database communicatively connected to the at least one processor; and
a memory storing executable instructions which, when executed, cause the at least one processor to perform operations including:
aggregating, from the at least one database, data associated with a plurality of entities, the aggregated data reflecting one or more relationships between two or more of the plurality of entities;
extracting, from the aggregated data, attribute information identifying loan amounts, property values, and appraisal sources;
converting the aggregated data into a knowledge graph database format;
populating one or more data structures with the extracted and converted attribute information;
generating a knowledge graph data structure having a plurality of subject nodes corresponding to the plurality of entities and a plurality of loan nodes corresponding to the extracted attribute information;
determining a geo-spatial neighborhood delineation using machine learning to classify a property's neighborhood membership based on the extracted attribute information, wherein the machine learning includes classification training through pruning noisy clusters through generalization;
updating, using machine learning analysis, the knowledge graph by changing an identified neighborhood for one or more homes represented as nodes on the knowledge graph;
generating a first statistical distribution of first attributes associated with a first appraisal source and a second statistical distribution of second attributes associated with a second appraisal source; and
detecting an anomaly in the first statistical distribution based on a comparison of the first statistical distribution and the second statistical distribution.