steel buildings damage classification by damage spectrum and decision tree algorithm

Results of damage prediction in buildings can be used as a useful tool for managing and decreasing seismic risk of earthquakes. in this study, damage spectrum and c4.5 decision tree algorithm were utilized for damage prediction in steel buildings during earthquakes. in order to prepare the damage spectrum, steel buildings were modeled as a singledegreeoffreedom (sdof) system and timehistory nonlinear analysis was carried out to develop a set of sdof structures. then, damage index was used to prepare the damage spectrum. data parameters required for training and evaluating the c4.5 decision tree algorithm were obtained from the results of damage spectra for steel structures and using krawinkler damage index also, two decision trees were trained based on quantitative indices. the first decision tree determined whether damage occurred in buildings or not and the second predicted severity of damage as repairable, beyond repair, or collapse. decision tree classification algorithm was used to predict damage to steel structures.