Housing price modeling using fixed and adaptive kernels in geographically weighted regression (A case study for district 5 of Tehran metropolitan city, Iran)

Housing is one of the main human needs in satisfying comfort and tranquility. In recent years, due to sharp changes in housing prices, especially in Tehran, estimating the price of apartments has become one of the most attractive topics among citizens. Fixed and adaptive kernels in geographically weighted regression have been used to analyze the spatial distribution of housing prices in District 5 of Tehran Municipality and the factors affecting it. The sample size includes 7464 apartment housing prices that were sold in 2018, 2019 and 2020. To compare the results of geographically weighted regression in each of the fixed and adaptive kernels, gaussian, exponential and bi-square functions have been used. The mean adjusted r-squared coefficient using gaussian, exponential and bi-square functions in fixed and adaptive kernels is 0.761 and 0.755, respectively. The results of this study show that due to the uniform distribution of data in the study area, the fixed kernel performs better than the adaptive kernel.

Housing price modeling using fixed and adaptive kernels in geographically weighted regression (A case study for district 5 of Tehran metropolitan city, Iran) Keywords:

geographically weighted regression , fixed kernel , adaptive kernel , housing prices , Tehran