In this paper, we mainly explore the effectiveness of two kernel-based methods, the convolution tree kernel and the shortest path dependency kernel, in which parsing information is directly applied to Chinese relation extraction on ACE 2007 corpus. Specifically, we explore the effect of different parse tree spans involved in convolution kernel for relation extraction. Besides, we experiment with composite kernels by combining the convolution kernel with feature-based kernels to study the complementary effects between tree kernel and flat kernels. For the shortest path dependency kernel, we improve it by replacing the strict same length requirement with finding the longest common subsequences between two shortest dependency paths. Experiments show kernel-based methods are effective for Chinese relation extraction.
展开▼
