Façade deterioration prediction with the use of machine learning methods, based on objective parameters and e-participation data

Abstract

Condition monitoring and timely repair of residential buildings is an important task when ensuring a comfortable life in cities. In the case of large metropolitan areas, it is a difficult task to perform continuous objective condition monitoring for tens of thousands of residential buildings by efforts of experts. However, residential infrastructure health can be predicted on the basis of indirect data. These can be objective building parameters or subjective data on citizens’ complaints about deterioration. In cities today, it is possible to collect such data in machine-readable form from various information systems. This article proposes a method to predict external deterioration of buildings on the basis of indirect data, using machine learning and SMILE Low-coding platform. Based on the results of method approbation, which used data of a metropolis, the significance of electronic participation data and objective parameters of objects for façade deterioration forecast was assessed. Options for further research are proposed to improve the quality of deterioration predicting by using data on citizens’ complaints about infrastructure damage.