TiMoS

该数据集名为TiMoS，包含了5623部电影剧情概要以及从类似维基百科的数据库TVTropes收集的95种不同剧情模式。每种剧情模式都附有定义和人工标注的分类。数据集被划分为训练集（占72.19%）、验证集（占12.72%）和测试集（占15.09%）。这些剧情模式被归类为四大类：角色特质、角色互动、情境和故事情节。规模上，该数据集涵盖了5623部电影剧情概要和95种剧情模式，其任务是进行电影剧情模式的检测。

The dataset named TiMoS contains 5,623 movie plot summaries and 95 distinct narrative tropes collected from the wiki-like database TVTropes. Each trope is accompanied by its definition and manually annotated classifications. The entire dataset is split into three subsets: the training set (72.19%), the validation set (12.72%), and the test set (15.09%). These narrative tropes are categorized into four major groups: character traits, character interactions, situational contexts, and story plots. The core task of this dataset is the detection of movie narrative tropes.