Representation Matters: Measuring Model-Market Representation, a Heuristic to Simplify AI Complexity
Start Date
4-12-2024 3:50 PM
End Date
4-12-2024 5:20 PM
Description
Whenever we use sales to measure assessment performance or to estimate value, there is an assumption that properties that sell adequately and proportionally represent unsold properties. This session presents two methods for examining this assumption, measuring the degree of generalization in a regression model, identifying properties that are not represented by sales and directly comparing performance for sold and unsold properties.
Publication Date
December 2024
Recommended Citation
Williams, James; Keene, Kevin; and Raju, Alex, "Representation Matters: Measuring Model-Market Representation, a Heuristic to Simplify AI Complexity" (2024). International Research Symposium. 8.
https://researchexchange.iaao.org/irs/irs24/sessions/8
Representation Matters: Measuring Model-Market Representation, a Heuristic to Simplify AI Complexity
Whenever we use sales to measure assessment performance or to estimate value, there is an assumption that properties that sell adequately and proportionally represent unsold properties. This session presents two methods for examining this assumption, measuring the degree of generalization in a regression model, identifying properties that are not represented by sales and directly comparing performance for sold and unsold properties.
