CIPCaD-Bench
收藏资源简介:
CIPCaD-Bench是一个专为因果发现方法设计的连续工业过程数据集。该数据集由意大利教育、大学和研究部支持,包含两个公开数据集:一个来自Tennessee Eastman模拟器的故障检测和过程控制数据集,另一个来自超加工食品制造厂的数据集,包含工厂描述和多个基本事实。这些数据集用于提出基于不同度量和CD算法的基准测试程序。CIPCaD-Bench旨在测试CD方法在现实条件下的性能,以选择最适合特定目标应用的方法。数据集的应用领域包括工业过程监控、故障诊断和战略决策支持。
CIPCaD-Bench is a continuous industrial process dataset specifically tailored for causal discovery methods. Supported by the Italian Ministry of Education, University and Research, this dataset suite includes two public datasets: one is the fault detection and process control dataset derived from the Tennessee Eastman simulator, and the other is sourced from an ultra-processed food manufacturing plant, which contains plant descriptions and multiple ground truths. These datasets are utilized to develop benchmarking procedures based on various metrics and causal discovery (CD) algorithms. CIPCaD-Bench is designed to evaluate the performance of CD methods under realistic scenarios, so as to select the most appropriate method for specific target applications. The application fields of this dataset cover industrial process monitoring, fault diagnosis and strategic decision support.




