Pattern Recognition Techniques Applied to Evaluation Engineering Problem

Free registration required

Executive Summary

The purpose of this paper is to present a Pattern Recognition methodology composed by Multivariate Statistical Analysis techniques, in order to build a Multiple Linear Regression statistical model to evaluate real estates according to their characteristics (variables, attributes). First, a Clustering Analysis was applied to the data of each urban estate class (apartments, houses or plots) to obtain homogeneous clusters within each class. Next, the Principal Components Analysis (P.C.A.) was applied to solve the multicollinearity problem that may exist among the variables in the model. The scores of the principal components are then the new independent variables and with them, the Multiple Linear Regression model was adjusted for each cluster of similar estates, within each class.

  • Format: PDF
  • Size: 134.5 KB