Abstract
Research background: Mass valuation is a process in which many properties are valued simultaneously with a uniform approach. An example of a procedure used for mass real estate valuation is the Szczecin Algorithm of Real Estate Mass Appraisal (SAREMA), which can be developed into a multiple regression model. The algorithm is based on a set of drawn representative properties. This set determines, inter alia, the quality of obtained valuations.
Purpose: The objective of the study is to verify the hypothesis whether changing the method of sampling representative properties from the originally used simple random sampling to stratified sampling improves the results of the SAREMA econometric variant.
Research methodology: The article presents a study that uses two methods of representative properties sampling – simple random sampling and stratified sampling. Errors of the models of valuation created taking into account both methods of sampling and different number of representative properties are compared. A key aspect of the survey is the choice of a better sampling method.
Results: The study has shown that stratified sampling improves valuation results and, more specifically, allows for lower root mean square errors. Stratified sampling yielded better results in the initial phase of the study with more observations, but reducing the percentage of strata participating in the draws, despite the increase in RMSE, guaranteed lower errors than the corresponding results based on simple sampling in all variants of the study.
Novelty: The article confirms the possibility of improving the results of mass property valuation by changing the scheme of representative properties sampling. The results allowed for the conclusion that stratified sampling is a better way of creating a set of representative properties.
Parameter Estimation Based on Double Ranked Set Samples with Applications to Weibull Distribution
