Serendipity Movie Test Data

Serendipity Movie Test Data The serendipity relatedness problem consists of, given an entity, to find another entity that is serendipitously related to it. The strategy adopted consists of diving the dataset in partitions based in a global feature, genre, and linking entities from different partitions according to connectivity criteria. This dataset supports the evaluation of approaches that address the serendipity relatedness problem. It covers the movie domain and uses data available in DBpedia and LinkedMDB, which are popular reference linked data datasets for this domain. The dataset contains 404 entities of the movies domain split in 15 cinematographic genres. For every entity, the dataset contains: (1) its linked data URIs; (2) its linked data datagraph, encompassing the network of entities that surrounds it; (3) a list of serendipitously suggested movies. And for every entity pair, the dataset contains: (1) the linked data paths that connect the pair; (2) a score according to the extracted paths. The data is compressed in .zip format and can be uncompressed by standard compression utilities. The data are split into two archives: dataset.zip: contains raw data (.json, .ttl). The underlying data and code can be accessed through standard text edit software.

惊喜关联（Serendipity）电影测试数据集

惊喜关联（Serendipity）问题的核心定义为：给定一个实体，寻找到与之存在偶然关联的另一实体。本数据集所采用的构建策略为：基于全局特征「类型（Genre）」对数据集进行分区，并依据连通性准则连接不同分区中的实体。
该数据集可用于支撑针对惊喜关联问题的各类算法方案的评估工作，其覆盖电影领域，采用DBpedia与LinkedMDB中的公开数据——二者均为该领域内广受认可的关联数据参考数据集。
本数据集共包含404个电影领域实体，这些实体被划分为15个电影类型。针对每个实体，数据集提供以下三类信息：(1) 其关联数据URI（Linked Data URIs）；(2) 其关联数据图（Linked Data Datagraph），涵盖该实体周边的实体关联网络；(3) 一份惊喜关联推荐电影列表。
针对每一对实体，数据集提供以下两类信息：(1) 连接该实体对的关联数据路径；(2) 基于提取路径计算得到的关联评分。
数据集以.zip格式压缩存储，可通过标准解压工具完成解压缩。数据集分为两个压缩包：dataset.zip 内含原始数据（格式为.json、.ttl），其底层数据与代码可通过标准文本编辑软件直接查看与编辑。