Automated Property Description Classification Using Numeric Representation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Manual review is required to determine the correct legal description among multiple legal descriptions for a property, which is time-consuming and inefficient, even with automated processes providing initial estimates.

Innovation Solution

A method that retrieves structured property descriptions from data sources, generates numeric representations using a trained model, conducts keyword searches, and combines results to compute the most likely classification, and uses numeric comparisons and textual centrality measures to automatically select the best description from multiple sources.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If automated processes are used to provide initial estimates of property descriptions, then productivity is improved, but reliability deteriorates due to inability to accurately determine the correct legal description among multiple descriptions

Engineering Contradiction:

Improveefficiency in property description processingVSAvoidaccuracy of legal description selection

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent introduces an automated classification system as an intermediary between multiple data sources and the final property description selection. The system uses trained models to generate numeric representations of descriptions, performs keyword searches, combines results, and computes classifications to objectively determine the most accurate legal description, eliminating the need for manual review while ensuring reliability

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transforms unstructured property descriptions into structured numeric representations using trained models. By converting text descriptions into comparable numeric formats and using keyword presence/absence values, the system enables automated comparison and classification, allowing computers to reliably select the correct description without manual intervention

Inventive Principle:

Principle #35Parameter changes

2Reliability

If manual review is performed to determine the correct legal description, then reliability is improved, but productivity deteriorates due to time-consuming processes

Engineering Contradiction:

Improveaccuracy of legal description determinationVSAvoidtime efficiency in title product creation

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent implements a self-service automated system that independently retrieves property descriptions from multiple data sources, generates numeric representations, performs keyword searches, combines results, computes classifications, and selects the most accurate description without requiring human intervention. The system serves itself by automatically determining the correct legal description with high reliability

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent performs preliminary actions by pre-training models on structured descriptions and pre-establishing keyword lists before actual property description processing. This preliminary preparation enables the system to quickly and accurately classify new descriptions without manual review, maintaining high reliability while improving productivity

Inventive Principle:

Principle #10Preliminary action

3Loss of information

If multiple data sources are searched for property descriptions, then completeness of information is improved, but device complexity increases due to need to process and compare multiple descriptions

Engineering Contradiction:

Improvecompleteness of property description dataVSAvoidcomplexity of processing multiple descriptions

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent segments the complex task of selecting the correct property description into distinct processing stages: retrieving descriptions from multiple data sources, generating numeric representations using trained models, performing keyword searches, combining results, and computing classifications. This segmentation simplifies the overall process by breaking it into manageable, automated steps that can be systematically executed

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent transforms diverse property descriptions from multiple data sources into a unified numeric representation format. By converting all descriptions into comparable numeric vectors and using standardized keyword presence/absence values, the system simplifies the complexity of processing multiple different description formats while maintaining completeness of information

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11232114B1System and method for automated classification of structured property description extracted from data source using numeric representation and keyword search

Publication Date: 2022.01.25 FIRST AMERICAN FINANCIAL CORP

AI summary

Some implementations of the disclosure are directed to retrieving a property description describing a real property; generating a numeric representation of the property description; determining one or more keywords of a keyword list present or absent in the property description; generating one or more values corresponding to a presence or absence of the one or more keywords in the property description; combining the numeric representation of the property description with the one or more values to create a combination; and computing, based on the combination, a most likely classification of the property description from a plurality of classifications.