A novel architecture to identify locations for Real Estate Investment

The identification of a favorable location for investment is a key aspect influencing the real estate market of a smart city. The number of factors that influence the identification easily runs into a few hundreds (including floor space area, crime in the locality and so on). Existing literature predominantly focuses on the analysis of price trends in a given location. This paper aims to develop a set of tools to compute an optimal location for investment, a problem which has received little attention in the literature (analysis of house price trends has received more attention). In previous work the authors proposed a machine learning approach for computing optimal locations. There are two main issues with the previous work. All real estate factors were assumed to be independent and identically distributed random variables. To address this, in the current paper we propose a network structure to derive the relational inferences between the factors. However, solving the location identification problem using only a network incurs computational burden. Hence, the machine learning layers from the previous work is combined with a network layer for computing an optimal location with proven lower computational cost. A second issue is that the computations are performed on an online database which has inherent privacy risks. The online data, user information and the algorithms can be tampered through privacy breaches. We present a privacy preservation technique to protect the algorithms, and use blockchains to secure the identity of the user. This paper presents solutions to two interesting problems in the analysis of real estate networks: a) to design tools that can identify an optimal location for investment and b) to preserve the privacy of the entire process using privacy preserving techniques and block chains.

Automated summary

This paper aims to develop a set of tools to compute an optimal location for investment by establishing a network structure to derive relational inferences between real estate factors. To reduce computational burden, machine learning layers are combined with a network layer. Additionally, privacy preservation techniques are proposed to protect algorithms and blockchains are used to secure user identity.