Turkish Sentence Dataset for Word Partitioning

The dataset contains the selected sentences for the appearing words in the datasets of analogy, NER, POS, and Sentiment analysis. Each dataset is also given in the folder. The main.txt is the selected sentences that represent the words in the dataset. The dataset is named as train.txt for extrinsic tasks and sentence-tr.json for analogy task. The analogy task contains all the word analogies in JSON format.

本数据集包含针对类比、命名实体识别（Named Entity Recognition，NER）、词性标注（Part-of-Speech Tagging，POS）与情感分析四类任务数据集内出现词汇所遴选的语句。各类任务的对应数据集均已存放于文件夹中。文件main.txt中存储了用于表征该数据集内目标词汇的遴选语句。针对外部任务的数据集命名为train.txt，类比任务的数据集则命名为sentence-tr.json。类比任务的全部词汇类比关系均以JSON格式存储。