francescoocurcio/ncbi-disease
收藏资源简介:
该数据集是一个用于命名实体识别(NER)任务的文本数据集,专注于疾病实体的标注。数据集中每条样本包含id、tokens(分词后的文本序列)和ner_tags(对应的命名实体标签)三个字段。标签体系采用BIO标注方案,包括O(非疾病实体)、B-Disease(疾病实体起始)和I-Disease(疾病实体内部)。数据集划分为训练集(5432个样本)、验证集(923个样本)和测试集(940个样本),总大小约3.19MB。
This dataset is a text dataset for Named Entity Recognition (NER) tasks, focusing on disease entity annotation. Each sample in the dataset includes three fields: id, tokens (tokenized text sequence), and ner_tags (corresponding named entity labels). The labeling scheme uses the BIO format, with categories including O (non-disease entity), B-Disease (beginning of disease entity), and I-Disease (inside of disease entity). The dataset is split into training set (5,432 samples), validation set (923 samples), and test set (940 samples), with a total size of approximately 3.19MB.




