Federal Home Loan Mortgage Corporation (freddie Mac)

Freddie Home Loan Mortgage Corporation (freddie Mac)

Fabio Quijada

Yuang Tang

Jianglong Li

Stanislav Kornienko

Dylan Nielson

Spyridon Kontogiorgis

Andrew Leister

Angela Rae Macias Helgeson

Matthew Vincent

Evelyn Suzette Johnson

Brendan McKinley

In an illustrative embodiment, systems and methods include storing, to feature vectors of real estate properties, both property attributes and at least one geospatial feature located within a predetermined distance of the property's location, evaluating, using machine learning (ML) algorithm(s), subsets of properties to identify impactful features having a significant impact on property value for each property subset, and, using an ML model trained for the impactful features of a given property subset, analyze feature data to obtain a) a set of similar properties and/or b) a property valuation for each of one or more real estate properties.

Systems and methods of establishing correlative relationships between geospatial data features in feature vectors representing property locations

Systems and methods are disclosed for automated rule-based income validation using electronic transaction data. One or more processors receive transaction data associated with a financial account, including data elements representing income transactions. Characteristics of the data elements are analyzed, and an income type label and an income source label are associated with the data elements. A cluster of data elements is identified as associated with a first income source label, the identified cluster including data elements associated with a first income type label. The cluster of data elements is verified, and data associated with the verified cluster is analyzed to determine one or more income streams. Based on a transaction rule associated with the one or more income streams, a composite income is calculated, and a net income associated with the first income type is determined. Based on the determined net income, a gross income distribution is generated.

Automated rule-based electronic transaction data analysis

In an illustrative embodiment, an automated system engineers customized feature vectors from geospatial information system (GIS) metadata. The system may include computing systems and devices for extracting metadata for GIS features located within a predetermined distance of a property from a GIS map file and storing the extracted GIS features within a feature vector. The system can augment each of the extracted GIS features with amplifying data features extracted from external data sources. The system can calculate a distance between the property and each extracted GIS feature, which establishes a relationship between the property and each GIS feature and associated amplifying data features. Amounts of correlation between each of the extracted GIS features and associated amplifying data features within the feature vector and a market assessment of the property location can be identified using a data model trained with a data set customized to characteristics of the property.

In an illustrative embodiment, an automated system assesses risks associated with real estate entities. The system may include computing systems and devices for extracting data attributes for real estate entities from received source data, and the extracted data attributes may each be associated with a particular data level for a real estate entity. The system can link data attributes in adjacent data levels with unique linkage information and compute metrics for each of the real estate entities from the data attributes where each computed metric may be associated with one of the data levels. The system can assess an amount of risk associated with each of the real estate entities based on the computed metrics and present the assessed amount of risk for one or more real estate entities to a remote computing device of a user responsive to receiving a risk assessment request.

Systems and methods of assessing viability of real estate entities

This disclosure relates to knowledge generation and implementation. A knowledge graph system comprises at least one processor, at least one database communicatively connected to the at least one processor, and a memory storing executable instructions. When executed, the instructions cause the at least one processor to aggregate data associated with a plurality of entities, the aggregated data reflecting one or more relationships between two or more of the plurality of entities. Attribute data identifying loan amounts, property values, and appraisal sources may be extracted from the aggregated data. A knowledge graph data structure may be generated having a plurality of subject notes corresponding to the extracted attribute information. Statistical distributions of attributes associated with one or more appraisal sources may be generated and an anomaly in a first statistical distribution may be detected based on a comparison of the first statistical distribution with a second statistical distribution.

Growing community of inventors

Reston, VA, United States of America

Fabio Quijada