Page 16 - Regression Guideline for AMC

Understanding regression mechanics
• When a factor is not significant, this means it did not reliably influence the sales prices of the modeled proper6es. This could be because there was not enough varia6on among the proper6es on non-­‐significant characteris6cs, the varia6on in sales price that they “explain” was associated with other property characteris6cs in the model, or market valua6on was not responsive to these characteris6cs.
• Based on the sta6s6cally significant property characteris6cs, the model produced a set of regression weights (the B coefficients in the equa6on). The table below shows the characteris6cs and associated regression weights. Nega6ve weights mean property value decreases with increasing values of that characteris6c:
Characteris:c
B
Characteris:c
B
Bank Owned (yes/no)
-­‐44,529.60
Number of bathrooms
46,820.8 3
Number of bedrooms
-­‐32,781.27
Square footage
153.08
Number of stories
-­‐40,025.42
Pool (yes/no)
48,435.9 1
Year of construc6on
1,553.22
16