Using multiple features and statistical model to calculate text units similarity | IEEE Conference Publication | IEEE Xplore