Mining real estate listings using Oracle data warehousing and predictive regression

The object of this paper is to introduce our development experience of a data mining system for prospective real estate sellers and buyers to determine properties price. The prediction of continuous values of properties selling prices is modeled by a statistical technique called predictive regression. The prerequisite of this data mining process is to design a data warehouse that contains a wide variety of real estate listings in the related areas. The data source is extracted from multiple listing services (MLS) database. It is cleansed and transformed at a staging area. The data warehouse design for this system is a star schema with one large fact table surrounded by three dimension tables. Loading data into the warehouse is the final step in creating data warehouse as preparation for data mining. Oracle data warehousing tool kits were used in the data warehouse construction. Visual Basic .Net was used to implement the data mining system.