Accuracy measurement of Random Forests and Linear Regression for mass appraisal models that estimate the prices of residential apartments in Nicosia, Cyprus

Abstract

Abstract. The purpose of this article is to examine the prediction accuracy of the Random Forests, a machine learning method, when it is applied for residential mass appraisals in the city of Nicosia, Cyprus. The analysis is performed using transaction sales data from the Cyprus Department of Lands and Surveys, the Consumer Price Index of Cyprus from the Cyprus Statistical Service and the Central Bank of Cyprus' Residential Index (Price index for apartments). The Consumer Price Index and the price index for apartments record quarterly price changes, while the dependent variables for the computational models were the Declared and the Accepted Prices that were conditional on observed values of a variety of independent variables. The Random Forests method exhibited enhanced prediction accuracy, especially for the models that comprised of a sufficient number of independent variables, indicating the method as prominent, although it has not yet been utilized adequately for mass appraisals.

Evaluation of Light Gradient Boosted Machine Learning Technique in Large Scale Land Use and Land Cover Classification

“…However, when compared to simple vector machines, the Random Forests algorithm proved to more stable and robust. These findings reflect those of other authors, who argued that Random Forests could be more accurate as long as the model is designed to specify random feature subsets [17].…”

A Fully Automated Adjustment of Ensemble Methods in Machine Learning for Modeling Complex Real Estate Systems

“…Reviewing the literature available so far, it can be concluded that the results obtained in terms of MAPE are better than those obtained in [26] with a value of 25.2% in Nicosia or in [46] for the US with 20.9%. e results are similar to those obtained in [14] with a value that ranges from 19.02% to 15.89% in Madrid and worse than those obtained for Ljubljana in [24] with an average MAPE of 7.28.…”

“…In addition, the conclusion emphasizes that it is not necessary to change the variables used in each county and that the accuracy of the model is practically the same using a series of common attributes for all of them. Dimopoulos et al [26] develop an application to compare the behaviour of random forest and linear regression in estimating the prices of residential apartments in Nicosia (Cyprus). e results verify that the best behaviour in predictive terms is that of random forest, with average MAPE values of 25.2%.…”

“…ese machine learning methods have been applied to the estimation of real estate properties in very specific locations. Research has been carried out by Jaen [12] in Coral Gables (Florida, USA); Fan et al [13] in Singapore (Republic of Singapore);Özsoy and Şahin [23] in Istanbul (Turkey); Del Cacho [14] in Madrid (Spain); Pow et al [17] in Montreal (Canada); Ceh et al [24] in Ljubljana, (Slovenia); Nguyen [25] in 5 counties of USA; and Dimopoulos et al [26] in Nicosia (Cyprus). In contrast, Pérez-Rave et al [20] deal with the estimation of dwelling prices for a whole country, Colombia, using independent variables identifying the city in which each property is located, thus proposing a unique model for the entire country (with a sample of 61,826 properties).…”

Preface: Earth surveillance and space-based monitoring of the environment: integrated approaches

“…Based on timeseries analysis of multispectral Sentinel-2 datasets, Evagorou et al (2019) obtained bathymetric data for shallow waters, (up to 30 m below sea level), using freely and open distributed optical satellite images. The ratio transform algorithm was implemented for twelve (12) monthly images covering thus a whole year.…”