Page 36 - Regression Guideline for AMC

R-­‐Square
• The variance that remains unexplained by the regression model can be due to errors in the measurement of the variables in the model (e.g., proper6es that do not have accurate measurement of gross living area, number of bedrooms, etc.), to sales anomalies such as a property being sold for well below its value owing to an estate sale, or to characteris6cs not included in the model because they were not available in the MLS data used for the analysis. This again, highlights the importance of careful data screening and selec6on of proper6es for analysis.
• R-­‐square is the most commonly used metric of regression model fit but not the only one. Because it’s value is affected by both the strength of associa6on as well as predic6on accuracy, other sta6s6cs are used to evaluate model fit and, par6cularly, the accuracy of the es6mated values. One such sta6s6c that is becoming more common is called the Mean Absolute Percent Error (MAPE).
36