House Price Prediction Research Articles

Local Spatiotemporal Modeling of House Prices: A Mixed Model Approach

The real estate market has long provided an active application area for spatial–temporal modeling and analysis and it is well known that house prices tend to be not only spatially but also temporally correlated. In the spatial dimension, nearby properties tend to have similar values because they share similar characteristics, but house prices tend to vary over space due to differences in these characteristics. In the temporal dimension, current house prices tend to be based on property values from previous years and in the spatial–temporal dimension, the properties on which current prices are based tend to be in close spatial proximity. To date, however, most research on house prices has adopted either a spatial perspective or a temporal one; relatively little effort has been devoted to situations where both spatial and temporal effects coexist. Using ten years of house price data in Fife, Scotland (2003–2012), this research applies a mixed model approach, semiparametric geographically weighted regression (GWR), to explore, model, and analyze the spatiotemporal variations in the relationships between house prices and associated determinants. The study demonstrates that the mixed modeling technique provides better results than standard approaches to predicting house prices by accounting for spatiotemporal relationships at both global and local scales.

Using machine learning algorithms for housing price prediction: The case of Fairfax County, Virginia housing data

House sales are determined based on the Standard & Poor’s Case-Shiller home price indices and the housing price index of the Office of Federal Housing Enterprise Oversight (OFHEO). These reflect the trends of the US housing market. In addition to these housing price indices, the development of a housing price prediction model can greatly assist in the prediction of future housing prices and the establishment of real estate policies. This study uses machine learning algorithms as a research methodology to develop a housing price prediction model. To improve the accuracy of housing price prediction, this paper analyzes the housing data of 5359 townhouses in Fairfax County, Virginia, gathered by the Multiple Listing Service (MLS) of the Metropolitan Regional Information Systems (MRIS). We develop a housing price prediction model based on machine learning algorithms such as C4.5, RIPPER, Naïve Bayesian, and AdaBoost and compare their classification accuracy performance. We then propose an improved housing price prediction model to assist a house seller or a real estate agent make better informed decisions based on house price valuation. The experiments demonstrate that the RIPPER algorithm, based on accuracy, consistently outperforms the other models in the performance of housing price prediction.

On Weighted Support Vector Regression

We propose a new type of weighted support vector regression (SVR), motivated by modeling local dependencies in time and space in prediction of house prices. The classic weights of the weighted SVR are added to the slack variables in the objective function (OF‐weights). This procedure directly shrinks the coefficient of each observation in the estimated functions; thus, it is widely used for minimizing influence of outliers. We propose to additionally add weights to the slack variables in the constraints (CF‐weights) and call the combination of weights the doubly weighted SVR. We illustrate the differences and similarities of the two types of weights by demonstrating the connection between the Least Absolute Shrinkage and Selection Operator (LASSO) and the SVR. We show that an SVR problem can be transformed to a LASSO problem plus a linear constraint and a box constraint. We demonstrate the capabilities of the doubly weighted approach through an example of prediction of house prices. The weight functions in the house pricing model depend on the geographical distance to the house of interest and the difference in time of sale (CF‐weights) as well as the differences lying in variables (OF‐weights), such as house size and number of floors. The results illustrate that the combination of the two types of weights describes the relative importance of observations very well and lowers the influence of possible outliers. Therefore, it enables the SVR models to have good performance. Copyright © 2014 John Wiley & Sons, Ltd.

Predicting house price via gene expression programming

PurposeThe housing sector is one of the main sources of economic growth in both developing and developed countries. Although many methods for modeling house prices have been proposed, each has its own limitations. The present paper aims to propose gene expression programming (GEP) as a new approach for prediction of housing price.Design/methodology/approachThis study introduces gene expression programming (GEP) as a new approach for predicting housing price. This is the first time that this metaheuristic method is used in the housing literature.FindingsThe housing price model based on the gene expression programming is compared with a least square regression model that is derived from a stepwise process. The results indicate that the GEP‐based model provides superior performance to the traditional regression.Originality/valueData used in this study is derived from the Household Income and Expenditure Survey (HIES) in Iran that is conducted by the Statistical Center of Iran (SCI). Housing price model is estimated by administering the questionnaires of this survey in Hamedan Province. To show the applicability of the derived model by GEP technique, it is verified applying parts of the data, namely test data sets that were not included in the modeling process.

Predicting housing prices according to expected future interest rate

In this article, I present a theoretical model for predicting future housing prices as a function of the expected future interest rate, housing depreciation and rent rate. Focusing on the notion of arbitrage, where the returns on investing in housing is equal to the return on bonds, I forecast housing prices in the US as a function of the expected future interest rate on corporate bonds graded at categories AAA or AA. I get that a change in the expected future bond yield will lead to a change in future housing prices.

A Coregionalized Model to Predict Housing Prices

In a uniequational hedonic model, the main source of spatial dependence is found in the explained variable, since the price of a house mainly depends on the housing prices in the neighborhood (although this can also be due to other factors, such as missing covariates and the model of choice). Dependence is one of the primary causes of spatial autocorrelation in disturbances. However, such disturbances may also be spatially correlated with the disturbances of other equations; in this case, they can be considered coregionalized. This paper presents a multi-equational hedonic regression model with coregionalized disturbances and heterotopic data. The model comprises two equations. The first explains housing prices using data from a sample, while the second explains an auxiliary variable, quality of the area, obtained from a different sample. The model is then applied practically to predict housing prices. [Key words: Cokriging, housing prices, geostatistics, multi-equational hedonic model, coregionalized.]

Research on House Price Prediction Based on Gray Markov Model

The paper mainly studies the house price prediction problem. Recently, with the rapid growth of domestic house price, it becomes conducive for the government to issue policy to regulate the house price in a visionary way to select a proper method to predict the price. At the same time, it is also very important to stabilize the domestic real estate market and promote the sound national economic development. The commercial residential housing price has the largest contribution rate among all the housing prices. Although commercial residential house price is featured with fluctuations, it basically shows the increasing tendency. Therefore, according to the existing tendency of the commercial residential house price, the paper uses the gray Markov Model for prediction. Gray GM (1,1)model will well simulate the long-term development tendency of the house price. In addition, Markov Model could remove the fluctuation of the house price data to increase the prediction accuracy.

Hybrid multilevel STAR models for hedonic house prices

This article proposes a procedure that exploits two heterogeneous data sets to derive spatial house price predictions for Austria in a semiparametric hierarchical regression framework. The first data set contains a large number of house price data, but only a small number of house characteristics (a long data set), giving rise to the omitted variable bias. In contrast, the second data set contains a small number of observations with a wide range of house price attributes (a wide data set). Although the latter allows for detailed estimation of the effects of house price characteristics, the sparse and uneven distribution over the research area leads to volatility in spatial effects. To circumvent these disadvantages, we propose a two-step approach: In the first step, we model the long data set, being aware of potential bias in the estimated effects. We apply a multilevel version of structured additive regression (STAR) models in order to account for nonlinearity in price functions as well as hierarchical spatial heterogeneity not explicitly modeled by covariates. The spatial prediction from this model is used as an explanatory covariate for the wide price data in the second stage. This novel modeling approach, a hybrid multilevel STAR model, avoids the bias from omitted variables on the one hand and yields robust spatial predictions on the other hand. We present detailed comparisons of nonlinear covariate effects and spatial heterogeneity from both stages, contrasting them with estimated effects from a reference model that is only based on the wide data set. The presented results support the approach taken in this paper, which proves particularly useful for spatial prediction of house prices.

The Impact of Technical and Non-technical Measures of Water Quality on Coastal Waterfront Property Values in South Florida

Due to the unprecedented population growth and increased economic activity in coastal areas, the health and hence value of coastal waterbody resources have been the subject of interest in recent years. In this article we estimate the value of a healthy waterbody through a hedonic property price analysis utilizing water quality as the amenity of interest. We compare hedonic analysis results using technical measures of water quality to the results using a non-technical measure of water quality “location grade” available in an urban coastal housing market of South Florida. Our results indicate that water quality improvement is associated with higher property values. In the comparison between technical and non-technical measures of water quality, we find that the technical measures provide better prediction of housing prices than the non-technical location grade. We further impute implicit prices for water quality improvement where significant mean willingness-to-pay (WTP) estimates range from $7,531 to $43,158.

Spatial Determinants of Housing Price Values in Istanbul

The aim of this study was to measure the effect of spatial characteristics on housing prices and to integrate an interpolation and regression model in terms of spatially predicting housing price values. In this paper, housing price is investigated by taking into consideration distance to city centre, transportation arteries and coasts, in addition to housing and neighbourhood characteristics as control variables. This investigation is conducted in two stages: firstly by the utilization of multiple regression analysis, and then by an interpolation technique which is generated to predict the spatial pattern of housing price on a continuous surface in order to test the reliability and consistency of the regression model. The results reveal that housing prices are significantly affected by spatial determinants referred to as the distance variables. By conducting a residual analysis from the regression model, housing price values are analysed and visualized in a continuous map which is globally consistent with the housing markets in Istanbul.

PERFORMANCE OF MULITPLE LINEAR REGRESSION AND NONLINEAR NEURAL NETWORKS AND FUZZY LOGIC TECHNIQUES IN MODELLING HOUSE PRICES

House price prediction continues to be important fo r government agencies insurance companies and real estate industry. This study investigates the perfor mance of house sales price models based on linear a nd non-linear approaches to study the effects of selec ted variables. Linear stepwise Multivariate Regress ion (MR) and nonlinear models of Neural Network (NN) and Adaptive Neuro-Fuzzy (ANFIS) are developed and compared. The GIS methods are used to integrate the data for the study area (Bathurst, Australia). While it was expected that the nonlinear methods wo uld be much better the analysis shows NN and ANFIS are only slightly better than MR suggesting q uestions about high R 2 often found in the literature. While structural data and macro-finance variables m ay contribute to higher R 2 performance comparison was the goal of this study and besides the Australi an data lacked structural elements. The results sho w that MR model could be improved. Also, the land value and location explained at best about 45% of the sale price variation. The analysis of price forecasts ( within the 10% range of the actual prediction) on average revealed that the non-linear models perform ed slightly better (29%) than the linear (26%). Th e inclusion of social data improves the MR prediction in most of the suburbs. The suburbs analysis shows the importance of socially based locations and also variance due to types of housing dominant. In general terms of R 2 , the NN model (0.45) performed only slightly bette r than ANFIS 0.39) and better than MR (0.37); but the linear MRsoc performed bett er (0.42). In suburb level, the NN model (7/15) performed better than ANFIS (3/15) but the linear M R (5/15) was better than ANFIS. The improved linear MR (6/15) performed nearly as well as the no n-linear NN. Linear methods appear to just as precise as the the more time consuming non linear methods in most cases for accounting for the differences and variation. However, when a much more in depth analysis is required non linear methods may prove to be more valuable. More research is nee ded in the area of house price modelling including more structural elements, modern buyer beliefs and the nature and type of risks noted in modern times.

Spatial and Temporal Dependence in House Price Prediction

This paper incorporates spatial and temporal dependence among housing transactions in predicting future house prices. We employ the spatiotemporal autoregressive model and structure the spatial and temporal weighting matrices as in Pace et al. (1998). We control for the time variation of both the attribute prices and the spatial and temporal dependence parameters through performing the analysis on an annual basis. Spatial heterogeneity is accounted for using experience-based definition of submarkets by real estate professionals. Using a comprehensive housing transaction data set from the Dutch Randstad region, we show that integrating the spatial and temporal dependence within the hedonic modeling improves the prediction power as compared to traditional hedonic model that neglects these effects.

Cycles in landed and non‐landed housing sub‐markets in Malaysia

PurposeThere exists a voluminous literature which examines house price diffusions across space and quality tiers. Numerous observers have pointed out that housing price dynamics can also be analysed in terms of house price diffusions across different housing sub‐markets (e.g. landed vs non‐landed housing). However, there has been no research to document these cyclical interactions. The purpose of this paper is to remedy this research gap by offering evidence of cyclical dynamics between different housing sub‐markets for the case of Malaysia.Design/methodology/approachThe dynamics between house prices of non‐landed housing (condominium) and landed housing (terrace, semi‐detached and detached housing sub‐markets) are analysed. Band‐pass filters are first used to extract the cyclical components from the four house price series. Next, interactions between house price cycles are documented using cross‐correlation analysis, Granger‐causality tests and impulse response functions. Finally, the out‐of‐sample forecasting performance of the respective house price series is examined.FindingsThe author finds that condominium price cycles are exogenous in that they lead the price cycles in other housing sub‐markets by one‐ to two‐quarters. Second, condominium price cycles drive price cycles in the terrace and detached sub‐markets. Third, cycles in condominium prices are able to predict house price cycles in other sub‐markets with a high degree of precision.Originality/valueThis is the first known study that establishes stylised facts of lead‐lag relationships between house price cycles in different housing sub‐markets. The findings have several practical implications for housing investors who intend to optimise investment decisions on housing purchases.

Land use regulations and the returns to low-income homeownership

I investigate the relationship between land use regulations and the house prices of urban low-income homeowners in the United States between 1998 and 2010. Because low-income homeowners generally hold a larger fraction of their wealth as home equity and are less likely to sustain homeownership, the overall financial position of this population is especially dependent on house price movements during short holding periods and the timing of house purchase and sale. The data used for analysis include zip-code-level house price estimates provided by Fannie Mae and proprietary mortgage origination data for a representative sample of low-income US homeowners who received community reinvestment mortgages between 1998 and 2004. I estimate a series of linear mixed-effects models to predict house price appreciation and the risk premium associated with investment in housing as opposed to a risk-free asset. Controlling for a variety of related factors, including land scarcity and subprime lending activity, I find that land use regulations are associated with greater house price volatility and a greater risk premium for urban low-income households. Thus, the results suggest that the dependence of low-income homeowner wealth on the timing of house purchase and sale is weaker in less regulated markets.

International Real Estate Review

This study evaluates the ability of a range of popular aggregate house price indexes to predict house prices out-of- sample at the transaction level for a small geographic area. The analysis particularly addresses the utility of spatial econometric methods. The results suggest that spatial econometric methods, which more explicitly consider the spatial aspects of observed house prices, provide better predictive accuracy as compared to more traditional estimation techniques, such as the repeat sales index, a hybrid repeat sales-hedonic price index, and hedonic price models estimated through least squares. The conclusions are drawn from a sample of 38,984 single-family residential real estate transactions for the city of Milwaukee, Wisconsin over the years 2002-2008.

The Economic Impact of Anticipated House Price Changes-Evidence from Home Sales

If realized house prices have the wealth effect and the collateral effect on the economy, anticipated house price changes should have similar economic effects. This article empirically analyzes the effects of single-family home sales, which are shown to be able to predict house prices in the literature, on economic production, using 372 metropolitan statistic areas in the United States from the first quarter of 1981 to the second quarter of 2008 in a panel vector error correction model. Changes in home sales are found to Granger because the growth rate of per-capita gross metropolitan product and the dynamic effects are visualized with impulse response functions. Supporting evidence for the economic impact of home sales is also found in contemporaneous regressions.

Recurring patterns in the run-up to house price busts

We present evidence that shows that large increases in credit and residential investment shares, along with deteriorating current account balances, provide useful leading indicators of house price busts. These variables also explain cross-sectional patterns in the build-up to the 2007 crisis. Interestingly, movements in output and inflation have little ability to predict house price busts.

Which factors capitalize into house prices? A Bayesian averaging approach

This paper investigates the robustness of 33 community-specific explanatory variables for house prices in the Swiss metropolitan area of Zurich using Bayesian model averaging. The analysis suggests a new way to perform hedonic variable selection and provides a minimal list of variables which may serve as a priori constraints when predicting house prices or estimating the effect of other community-specific characteristics in a metropolitan area in a highly developed country. In the context analyzed, the main variables which capitalize with a high posterior probability are location-specific real estate characteristics, municipal taxes and expenditure for culture, health and social well-being. Demographic as well as other socio-economic controls seem to be of minor importance.

Predicting House Prices with Spatial Dependence: A Comparison of Alternative Methods

This paper compares alternative methods for taking spatialdependence into account in house price prediction. We selecthedonic methods that have been reported in the literature toperform relatively well in terms of ex-sample predictionaccuracy. Because differences in performance may be due todifferences in data, we compare the methods using a single dataset. The estimation methods include simple OLS, a two-stageprocess incorporating nearest neighbors' residuals in the secondstage, geostatistical, and trend surface models. These models takeinto account submarkets by adding dummy variables or byestimating separate equations for each submarket. Based ondata for approximately 13,000 transactions from Louisville,Kentucky, we conclude that a geostatistical model withdisaggregated submarket variables performs best.

Delineating urban housing submarkets with fuzzy clustering

It has long been argued that the housing market is spatially compartmentalized within a metropolitan area. The argument has important implications for explaining how the housing market works—should the status quo be seen as an equilibrium state? Or if no equilibrium is reached, how do loosely interlaced submarkets function both independently and interdependently? We note that the body of literature has leaned toward testing the distinctiveness of housing submarkets given a priori housing submarkets. However, there seems to be a lack of interest in developing methods for deriving housing submarkets empirically. Fuzzy clustering is well suited to this problem, given that the boundary of housing submarkets is not often sharply delineated. The study applies a fuzzy c-means (FCM) algorithm to identify housing submarkets in the Buffalo–Niagara Falls region. The study is distinct from other FCM applications in three respects. First, we reflect on issues tied to choosing parameters of fuzzy clustering. Second, we introduce overlap measures to characterize the relationship between the clusters produced. Third, we evaluate the performance of fuzzy clustering in terms of hedonic prediction accuracy. Results show that stratified hedonic models predict house price better than a market-wide hedonic model. Fuzzy clustering solutions also yield better prediction, compared with hard clustering.