conll2000
收藏资源简介:
CoNLL-2000旨在将文本划分为在语法上相关的词语部分,用于组块分析任务,是CoNLL-2000的共享任务。它包含从华尔街日报语料库(WSJ)中划分的数据,其中训练数据包含8937个样本,测试数据包含2013个样本。每个样本包括id、tokens、词性标签(pos_tags)和组块标签(chunk_tags)等字段。CoNLL-2000提供文本组块分析所需的数据,但关于数据的来源、标注过程、授权许可等详细信息缺失。
CoNLL-2000, the shared task of CoNLL-2000, aims to segment texts into grammatically relevant word chunks for chunking analysis tasks. It consists of data extracted from the Wall Street Journal corpus (WSJ), with the training set containing 8937 samples and the test set including 2013 samples. Each sample covers fields such as id, tokens, part-of-speech tags (pos_tags), and chunk tags (chunk_tags). CoNLL-2000 provides the necessary data for text chunking analysis, but detailed information such as data sources, annotation procedures, and licensing terms is missing.




