Estimating the Technical Deterioration of Large-panel Residential Buildings Using Artificial Neural Networks☆

Abstract In order to identify the repair needs of large housing estates a simplified method for estimating the technical deterioration of large number of houses is needed. A method presented in the paper is based on the extracted data processed by means of artificial neural networks (ANN). The aim is to create the artificial neural network configurations for a set of data containing values of the technical deterioration and information about building repairs obtained in earlier years (or other information and building parameters) and next to analyze new buildings by the instructed neural network. The profit from using ANN is the reduction of the number of parameters. Instead of more than forty parameters describing a building, usually about ten are sufficient for satisfactory accuracy. Three types of ANN are used: MLP - Multilayer Perceptron, RBF - Radial Basis Function, SVM - Support Vector Machine. The paper presents the part of results of author's PhD dissertation [1] , [2] .