TV-ALP: A log dataset of television assembly line production under multi-person collaboration for process mining research

Process mining technology has been widely used to optimize the processes of various organizations, especially in enterprises. It facilitates cooperation between departments and prompts efficient process design and resource scheduling in the production workshop. However, as a data-driven approach, the lack of production logs hinders the development of enterprise process mining research. Therefore, we introduce a new benchmark dataset named TV-ALP to provide effective data support for the combination of process mining and workshop production. The dataset comes from the field survey experience and product operation instructions, which highly restores the operation of the production workshop and details the processing of television (TV) sets in the assembly line. We compare TV-ALP with other public datasets for detailed statistical analysis and provide benchmark performance for the remaining time prediction task. The experimental results show that TV-ALP can meet the requirements of process mining and analysis research in terms of both scale and quality. In addition, while maintaining the data commonality of public datasets, it also emphasizes the importance of role information and supports a series of role-based log studies. The complete dataset is accessible for download at https://github.com/Zzou-Sdust/TV-ALP-dataset.