Regression model for appraisal of real estate using recurrent neural network and boosting tree

Automated valuation model (AVM) is a mathematical program to estimate the market value of real estates based on the analysis of locations, neighborhood characteristics, and relevant property characteristics. The most common AVMs employed by the appraisal industry are based on multiple regression analysis. Other analytic tools such as statistical learning and fuzzy algorithms have become more popular because of the increasing capability of collecting a high volume of data and the advancement of machine learning. The new analytic model thus becomes possible to build a more sophisticated model to exploit the information embedded in the collected data. In this work, we proposed a boosting tree model facilitated with a Recurrent Neural Network (RNN) to forecast the average price of an area. The experimental results indicate that our model outperforms the existing models adopted in the appraisal industry.