Využití Street View a Družicových snímků pro odhad cen nemovitostí

Abstract

The advent of big data has opened many opportunities. One of which is its use in the real estate valuation industry. Big data sources replace common sources and statistical methods used to estimate property prices. At the same time, deep and machine learning is used using neural networks, in which hedonic and other models can be implemented. The main advantage of the hedonic model is the accurate prediction of real estate prices. Such an estimate serves as a tool for real estate industry, brokers and people who want to market their property. For individuals as well as real estate customers, the most important criterion is the exact price and they are satisfied with the use of black box models, such as neural networks. The work deals with the implementation of large image sources of street view data and satellite imagery into a hedonic model, using machine and deep learning, for estimating real estate prices.