PicTropes
收藏资源简介:
PicTropes数据集由DBTropes.org数据库提取,包含5925部电影及其使用的18,270个电影桥段。数据集大小庞大,平均每部电影包含43.434个桥段,每个桥段平均出现在14.086部电影中。创建过程涉及从DBTropes.org提取数据并转换为JSON格式,便于编程语言处理。数据集主要应用于机器学习和叙事生成研究,特别是用于生物启发技术在文学和开放世界视频游戏中生成背景故事的研究。
The PicTropes dataset is extracted from the DBTropes.org database, containing 5,925 films and 18,270 film tropes utilized across these works. Boasting a considerable scale, the dataset has an average of 43.434 tropes per film, with each trope appearing in an average of 14.086 films. Its creation involves extracting data from DBTropes.org and converting it into JSON format to facilitate processing by programming languages. The dataset is primarily utilized for research in machine learning and narrative generation, especially studies applying bio-inspired technologies to generate backstories in literary works and open-world video games.

- 1Overview of PicTropes, a film trope dataset巴德尔地图公司,旧金山,美国;格拉纳达大学和CITIC-UGR,西班牙;卡迪斯大学和CITIC-UGR,西班牙 · 2018年



