Bayesian Statistical Estimation in Real Estate Price Modeling: A Comparative Study with Traditional Regression

Abstract

Property valuation alongside financial planning and investment decision-making heavily depend on real estate price modeling as an essential process. Price estimation through Multiple Linear Regression (MLR) and Generalized Linear Models (GLM) is commonly practiced among experts however these methods struggle to represent complexities found in real estate markets. The research evaluates Bayesian statistical estimation as a cutting-edge substitute that optimizes predictions through previous distribution theory combined with MCMC simulation techniques. The research measures traditional and Bayesian regression models' performance by comparing them with BIC (Bayesian Information Criterion), AIC (Akaike Information Criterion) and R² measure. Prior market knowledge incorporated into Bayesian estimation leads to better model fitting performance along with improved predictive reliability according to the results. The research presents Bayesian modeling techniques as an enhancement for real estate price analysis which leads to better property market decision support.