Principles for Calculating AVM Performance Metrics
Abstract
An analysis of 5.3 million housing sales suggests that there are fundamental shortcomings with how AVM vendors currently calculate their AVM Performance Metrics, in particular the FSD.
We propose consistent methodologies, i.e., five best practice principles and a modified leave-one out, cross-validation technique, to calculate AVM Performance Metrics that comply with well established appraiser principles and allow a consistent evaluation and comparison of AVM performance. A research AVM, developed to estimate a specific target property’s market value from a dataset of comparable housing sales in Cedar Falls, Iowa, empirically illustrates that not
following these principles yields overly optimistic AVM Performance Metric values.
Keywords: AVMs; AVM Performance Metrics; Cross-Validation; Error Buckets; Failure Rate; FSD; Leave-One-Out; PRESS Statistic