MSAIRS

MSAIRS数据集由苏州大学研究团队创建，包含了从主流社交媒体平台提取的5000多条带有明确意图和贴图的聊天记录。数据集特别关注贴图对聊天情感分析和意图识别的影响，提供了同一文本与不同贴图配对的数据，以及包含相同图像但文本不同的贴图，以深入理解贴图在社交对话中的作用。该数据集适用于研究情感与意图的交互影响，特别是在难以直接通过语言表达情感和意图的社交场景中。

The MSAIRS dataset was created by a research team from Soochow University. It comprises over 5,000 chat records with explicit communicative intentions and attached stickers, extracted from mainstream social media platforms. This dataset specifically targets the impact of chat stickers on sentiment analysis and intention recognition in conversational contexts. It provides two categories of paired data: one where identical text is matched with distinct stickers, and another where the same sticker image is paired with varying textual content, to enable in-depth exploration of the role of stickers in social dialogues. This dataset is suitable for researching the interactive effects of sentiment and intention, particularly in social scenarios where emotions and intentions are difficult to convey directly through language alone.

搜集汇总

背景与挑战

背景概述

MSAIRS数据集由苏州大学团队构建，包含超过5000条来自主流社交媒体的聊天记录，每条记录均带有明确意图和贴图。该数据集通过提供同一文本与不同贴图的配对，以及相同图像与不同文本的组合，专门研究贴图对聊天情感分析和意图识别的影响，适用于探索情感与意图交互的复杂社交场景。