Machine Learning for Ratio Studies
Presentation Category
Modeling and Valuation
General Session Description
Ratio studies are an essential tool for evaluating the fairness and uniformity of assessments, but performing them thoroughly can be time consuming and limited by traditional spreadsheet methods. This session explores how machine learning can be applied to ratio study data to identify inequities and patterns that may not be visible through simple statistics. Participants will see how machine learning such as SHAP and recursive partitioning methods can detect areas or submarkets where ratios diverge, suggesting potential issues with valuation or data quality. The focus is on plain language interpretation and practical understanding rather than complex coding, showing how assessors can use open-source software to strengthen equity analysis and better direct revaluation resources.
Start Date
4-1-2026 9:00 AM
End Date
4-1-2026 10:00 AM
Machine Learning for Ratio Studies
Ratio studies are an essential tool for evaluating the fairness and uniformity of assessments, but performing them thoroughly can be time consuming and limited by traditional spreadsheet methods. This session explores how machine learning can be applied to ratio study data to identify inequities and patterns that may not be visible through simple statistics. Participants will see how machine learning such as SHAP and recursive partitioning methods can detect areas or submarkets where ratios diverge, suggesting potential issues with valuation or data quality. The focus is on plain language interpretation and practical understanding rather than complex coding, showing how assessors can use open-source software to strengthen equity analysis and better direct revaluation resources.