Page 4 - Regression Guideline for AMC

Evalua6ng Mul6ple Regression Model FiRng Real Estate Appraising
• A regression model will always provide es6mates that can be used to value the subject property. However, the ques6on is the accuracy of those es6mates and consequently the subject property’s appraised value.
• And, no ma.er how accurate the es6mates, regression models do not predict perfectly. There is always some predic6on error. We want to determine not if the regression model is perfect but if the margin of error associated with the predicted value for the subject property is reasonably accurate within some acceptable tolerance level (e.g., within +/– 5% of the “true” value).
• Given that a regression analysis can be either useful (accurate) or not useful (inaccurate) depending on how it is done and the quality of the data used, we need the ability to dis6nguish between accurate and inaccurate models. How to produce the most accurate model possible given the data and evaluate the accuracy of the model is the focus of this presenta6on/discussion.
4