Abstract
This paper proposes two methods to estimate the similarity of ground motions based on the probability density function of nonlinear response of strcutures with parametric uncertainty. One uses mutual information to estimate the difference of the probability density functions and the other Kullback. Leibler divergence. The performacne of the proposed methods are discussed by numerical simulations using artificially modified strong motion records. It was presented that the proposed methods rationally estimate the range of modification.
