Seismic oscillation record selection method based on clustering analysis

The invention relates to a seismic oscillation record selection method based on clustering analysis. The method comprises the following steps: 1, selecting seismic oscillation records based on a strong earthquake database; 2, selecting an earthquake magnitude and epicentral distance strip range, and giving specific earthquake information of each seismic oscillation record; 3, according to a fuzzy C-mean value clustering algorithm, taking the mean value of the magnitude and the epicentral distance of each seismic oscillation record as a clustering center, repeatedly modifying the clustering center and a data membership degree through an iteration process, and giving a preliminary trial calculation clustering analysis result; 4, introducing a simulated annealing algorithm to converge to global optimal clustering analysis; and 5, on the basis of carrying out clustering analysis according to the magnitude and the epicentral distance, giving a response spectrum curve of each group of seismic oscillation records un