MmMtCSD
收藏资源简介:
MmMtCSD数据集由深圳技术大学创建,旨在解决多模态多轮对话立场检测问题。该数据集包含21,340条标注数据,涵盖文本和图像两种模态,主要针对“特斯拉”和“比特币”两个目标进行立场标注。数据集通过Reddit平台收集,经过严格的前处理和多轮专家标注,确保数据的高质量和标注的一致性。该数据集的应用领域主要集中在社交媒体分析和立场检测研究,旨在通过多模态信息提升立场检测的准确性。
The MmMtCSD dataset was created by Shenzhen Technology University, aiming to address the task of multimodal multi-turn conversational stance detection. It contains 21,340 annotated samples, covering two modalities: text and image, with stance annotations primarily targeting two topics: "Tesla" and "Bitcoin". The dataset was collected from the Reddit platform, and underwent strict preprocessing and multi-round expert annotation to ensure high data quality and annotation consistency. Its application fields mainly focus on social media analysis and stance detection research, with the goal of improving the accuracy of stance detection by leveraging multimodal information.




