Abstract

The main purpose of the article is to investigate the formation of house prices in Slovenia with a combination of geographically weighted regression and hierarchical clustering. It also demonstrates how to operationalize the geographically weighted regression for house price forecasting, tailored to the regional market features. We analyzed data on house sales in Slovenia for 2015 by estimating a global and regional regression models. The effect of location on house value was described by employing house value zones and through regionalisation. Results reveal that location, railway proximity, access to gas network, house area, house age, and area of building and farming lot belonging to a house have a statistically significant effect on house prices.

Key words: analysis of house prices, geographically weighted regression, clustering, regionalisation