Data from other similar houses in your area were used to predict the price of your house by identifying a few variables which capture the variability in home prices. This was done by looking at the relationship between predictor variables (e.g., square feet, bathrooms, etc.) and home price, the response variable. A large sample size was gathered to account for representative homes of the area, but of course, there will always be exceptions and some degree of bias. Regression model building isn't an exact science, so there will always be a margin of error induced in these predictions. The less representative your house is compared to other houses in the area, the larger one could expect this margin of error to be.
Learn More