Property Embedding Rankings for Accurate Comparable Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing property valuation models require heavy computational resources and are biased due to manual selection of comparable properties, leading to inaccurate valuations.
Innovation Solution
A system that processes embeddings of properties and candidate comparable properties to learn an implicit ranking, using an implicit comparable-ranking model to select better comparable properties based on compatibility probabilities, reducing computational burden and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual selection of comparable properties is used, then appraisers can subjectively adjust sale prices to reflect differences, but the process is biased and time-consuming
Solution Approach 1:
The patent replaces the manual mechanical process of appraiser selection and adjustment with an automated machine learning system. The system uses trained models to automatically select comparable properties and compute adjustments, eliminating the time-consuming manual process while maintaining or improving accuracy through consistent algorithmic application.
Solution Approach 2:
The system enables self-service by allowing the valuation process to be performed automatically without human intervention. The machine learning models independently select comparable properties, compute adjustments based on learned patterns, and generate valuations, freeing appraisers from routine tasks while improving consistency.
2Measurement precision
If traditional property valuation models are used, then comprehensive analysis of comparable properties can be performed, but heavy computational resources are required
Solution Approach 1:
The patent applies preliminary action by pre-training machine learning models on historical property data before actual valuation. This pre-training phase learns patterns and relationships in advance, so that during actual valuation, the system can quickly select comparable properties and compute adjustments using the pre-learned knowledge, reducing real-time computational requirements while maintaining accuracy.
3Productivity
If statistical methods with historical transactions are used, then predicted values can be generated, but the models require extensive data processing and are complex
Solution Approach 1:
The patent extracts and separates the complex model training and selection processes from the actual valuation process. The machine learning models are trained separately on historical data to learn patterns, then this learned knowledge is applied during valuation. This separation simplifies the actual valuation operation while maintaining the productivity benefits of comprehensive historical analysis.
Data Source
AI summary
Comps processing systems and methods for processing embeddings of properties and candidate comparable properties thereof to learn an implicit ranking of the candidate comparable properties are disclosed. The comps processing system uses embeddings, predicted values, and actual values of particular properties and candidate comparable properties thereof to train an implicit comparable-ranking model to produce comparable-adjusted predicted values and intermediate outputs. The comps processing system can then extract the intermediate outputs from the trained implicit comparable-ranking model, when applied to subject properties. The intermediate outputs can include compatibility probabilities, which represent a degree of compatibility between the subject property and a candidate comparable property thereof. Accordingly, the comps processing system can learn the similarities between a subject property and candidate comparable properties thereof when learning to value the subject property. The compatibility probabilities can be used to rank the candidate comparable properties and select those most similar to the subject property.