A recurrent neural network for for real estate price estimation
DOI:
https://doi.org/10.31548/energiya5(69).2023.088Abstract
The paper considers the problem of estimating the price of real estate. Automation of the solution to the specified problem is one of the methods of providing an objective assessment, which excludes such subjective factors as an arithmetic calculation error, assessment under the influence of emotions, assessment under the influence of the pursuit of one's own goals. Most often, regression analysis (hedonic regression) and machine learning methods are used to solve the problem of estimating the price of real estate. The purpose of this research is to build a model for estimating the price of real estate not only on the basis of the usual quantitative indicators (for example, area, number of rooms, floor, etc.), but also on the basis of the textual description of the real estate. In this paper we consider a real estate in new residential builings in the Kyiv. To achieve the goal, a mathematical model was developed for the classification of real estate class based on a textual description using a recurrent neural network. The model is developed as a multi-layer feed-forward neural network that accepts textual data describing a new residential building and passes it through a series of hidden layers, where each layer consists of neurons. The simulated categorical variable 'predictedClass' was used as the independent variable of the linear regression to calculate the price of an apartment in a new building. Compared to classical linear regression with quantitative regressors, the new model provided a high R2 with a minimal number of variables.
Key words: real estate, categorical data, classification, recurrent neural network, linear regression
References
Sopranzetti, B.J. (2010). Hedonic Regression Analysis in Real Estate Markets: A Primer. In: Lee, CF., Lee, A.C., Lee, J. (eds) Handbook of Quantitative Finance and Risk Management. Springer, Boston, MA. https://doi.org/10.1007/978-0-387-77117-5_78
Vishwakarma, Vijay Kumar, Forecasting Real Estate Business: Empirical Evidence from the Canadian Market (2013). Global Journal of Business Research, v. 7 (3) pp. 1-14, 2013, Available at SSRN: https://ssrn.com/abstract=2148507
Baur, Katharina & Rosenfelder, Markus & Lutz, Bernhard. (2023). Automated real estate valuation with machine learning models using property descriptions. Expert Systems with Applications. 213. 119147. 10.1016/j.eswa.2022.119147.
Fan, Gang-Zhi & Ong, Seow Eng & Koh, Hian. (2006). Determinants of House Price: A Decision Tree Approach. Urban Studies. 43. 2301-2316. 10.1080/00420980600990928.
Wanga, Pei-Ying & Chen, Chiao-Ting & Su, Jain-Wun & Ting-Yun, Wang & Huang, Szu-Hao. (2021). Deep Learning Model for House Price Prediction Using Heterogeneous Data Analysis Along With Joint Self-Attention Mechanism. IEEE Access. PP. 1-1. 10.1109/ACCESS.2021.3071306.
Tymchuk, O., Pylypenko, A., Iepik, M. Forecasting of Categorical Time Series Using Computing with Words Model (Conference Paper) 9th International Scientific Conference ""Information Technology and Implementation"" Workshop, IT and I-WS 2022; Kyiv; Ukraine; CEUR Workshop Proceedings. Volume 3384, 2022, Pages 151-159 URL: http://www.scopus.com/inward/record.url?eid=2-s2.0-85159621875&partnerID=MN8TOARS
Downloads
Published
Issue
Section
License
Relationship between right holders and users shall be governed by the terms of the license Creative Commons Attribution – non-commercial – Distribution On Same Conditions 4.0 international (CC BY-NC-SA 4.0):https://creativecommons.org/licenses/by-nc-sa/4.0/deed.uk
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
