LG-AI-Research/SemEval-STM

SemEval-STM数据集基于SemEval数据构建，旨在展示分段主题建模（STM）的优越性。该数据集包含两种主题分配范式的注释：基于文档的主题分配（DBTA）和基于分段的主题分配（SBTA）。数据集包含多个配置和分割，详细描述了数据生成管道的每个步骤。此外，还提供了数据来源、许可证和目录结构的信息。

SemEval-STM is a dataset built on SemEval data to demonstrate the superiority of Segment Topic Modeling (STM). It contains annotations for two topic allocation paradigms: Document Based Topic Allocation (DBTA) and Segment Based Topic Allocation (SBTA). The dataset includes various configurations and splits, with detailed descriptions of each step in the data generation pipeline. It also provides information about the data source, license, and directory structure.