Classificação de preços de imóveis utilizando a técnica de floresta aleatória

Abstract

Real estate pricing is relevant to several issues, such as real estate investment, simulation and even urban planning. Therefore, investors, buyers and sellers of real estate are potentially interested people. Information gathering from electronic databases is nontrivial; this motivates the development of Knowledge Discovery in Databases (KDD). This process has, among its steps, the Data Mining (MD) and evaluation steps. To carry out mining, some algorithms are used, in this specific application, the Random Forest (RF). Forthe evaluation, a method must be used on the generated discovery, such as the confusion matrix. Price forecasting, in turn, can be a goal of KDD. Comparisons can be made using the model generated by the work. In this work, the modeling was performed with a set of geographic data, the final classification being a price range.