Hedonic house price models often have errors that are related to location, which can affect the accuracy of the results. Traditional methods to address this issue may not consider all factors properly. A new approach called the generalized additive model shows promise in providing more reliable estimates of spatial variables. This model is less restrictive and can help reduce the problem of spatial autocorrelation. However, challenges still exist when dealing with different types of spatial data. It is important to test the robustness of estimates using sensitivity analysis to ensure the accuracy of the results.