SCTB: A Chinese Treebank in Scientific Domain

Update History


SCTB is a phrase structure based Chinese treebank. The raw sentences are selected from the LCAS (National Science Library, Chinese Academy of Sciences) Chinese scientific paper corpus provided by Japan Science and Technology Agency (JST). Our annotation process follows that of the Penn Chinese treebank (CTB) with an exception of the "segmentation standard". We apply a Chinese word segmentation standard based on "short and consistent units" (Shen et al., 2016). The first version of release contains 5,133 sentences (138,781 words). The second version of release contains 12,175 sentences (328,562 words). This work is supported by the JST MT project "Project on Practical Implementation of Japanese to Chinese-Chinese to Japanese Machine Translation."


( (IP (NP (NT 近年)) (NP (NP (NN 稻瘟) (SFN 病)) (NP (NN 发生))) (VP (VV 呈) (NP (CP (IP (VP (VV 加重)))) 
(NP (NN 趋势)))) (PU 。)) )
( (IP (VP (VV 评价) (NP (NP (QP (CD 两)) (NP (NN 组))) (NP (NP (NN 疗效)) (CC 及) (NP (NN 安全) (SFN 性))))) 
(PU 。 )) )
( (IP (NP (NN 结果)) (VP (VV 表明) (PU :) (IP (NP (NP (NN 芝麻)) (NP (NP (DP (DT 全)) (NP (VV 生育) (SFN 期)))
(NP (NN 长短)))) (PU ,) (VP (PP (P 与) (NP (NN 温度))) (VP (VV 呈) (NP (PFA 正) (NN 相关)))))) (PU ;)) )


Copyright (C) Kurohashi-Kawahara Lab. and JST. You can use all the data under the terms of the Creative Commons Attribution 3.0 Unported license.


Please fill this form for download.


Contact and Bug Report