An Overview of the Forecasting Methods Used in Real Estate Housing Price Modelling

Abstract

Forecasting is very fundamental in real estate where the past transactions become the evidences while decision making for the present and the future. Several techniques and validation approached that were commonly used in housing price index forecasting. Beside the appropriate forecasting method, error calculation is one of the critical constraints in accuracy out of all methods. This paper overview the available methods and the types of error being considered in forecasting techniques. Then the forecasting methods, namely Multiple Regression Analysis (MRA) and Artificial Neural Network which are highly applied in forecasting modelling are compared over its error accuracy.

Predictive Accuracy of Logit Regression for Data-Scarce Developing Markets: A Nigeria and South Africa Study

“…The strength and similarity of the model capacity in both countries showed that each market signal correctly predicts turning points in the economy for as much as 75% (Nigeria) and 80% (South Africa) of the time. The misclassification rate for the Nigerian logistic model is, however, 5% higher which is similar to the average model error margins observed by Munusamy, Muthuveerappan and Baba [14]. Meanwhile, the classification accuracy of the South African logit model is higher than that of the Nigerian logit model.…”

“…A lot of research has been carried out to promote macroeconomic modelling as a means of improving forecasting accuracy. Some popular works in these subject areas include studies such as those of Munusamy, Muthuveerappan and Baba [14]; Jadevicius, Sloan and Brown [15]; Tsolacos [16]; Jadevicius, Sloan and Brown [17]; and Boshoff [4]. Buehler and Almeida [18] discussed the investment market in the United States relative to predicting commercial market bubbles as a part of decision making for global investors.…”

“…Munusamy, Muthuveerappan and Baba [14] considered a variety of literature regarding modelling types, accuracy and adoption of statistical modelling techniques. They reported that Multiple Regression Analysis (MRA) and Artificial Neural Networks (ANNs) were most widely used.…”