Publicação
Enhancing Healthcare Process Model Discovery Through Duplicate Task Identification
| Resumo: | Healthcare plays an increasingly vital role in our daily lives. Modern Hospital Information Systems (HISs) record and store detailed medical treatment processes for all patients in the form of event logs. Leveraging these logs, process mining techniques have been widely applied to extract valuable insights, optimize medical processes, and enhance healthcare service delivery. However, existing process discovery techniques struggle to effectively handle healthcare event logs, as these processes often involve a high frequency of duplicate tasks. To address these challenges, we propose a novel duplicate task-aware process discovery technique for healthcare. More specifically, it first analyzes contextual relations among tasks to identify the sequence of duplicate tasks. Then, duplicate tasks are relabeled using a transition system derived from the input event log. Finally, a directly-follows graph is generated based on transition adjacency relations and transformed into a Petri net using the Inductive Miner. The proposed technique is fully implemented in the opensource process mining platform ProM and evaluated using a public healthcare process case with six event logs. Comparative analysis with state-of-the-art discovery techniques demonstrates that our approach accurately identifies duplicate tasks and produces high-quality process models, achieving superior replay precision and reduced complexity. |
|---|---|
| Autores principais: | Su, Xuan |
| Outros Autores: | Liu, Cong; Lu, Faming; Cheng, Long; Zeng, Qingtian; Zhou, Jiehan |
| Assunto: | Duplicate Healthcare process Hospital information systems Model discovery Process mining tasks Information Systems Computer Science Applications Computer Networks and Communications Information Systems and Management Artificial Intelligence |
| Ano: | 2025 |
| País: | Portugal |
| Tipo de documento: | documento de conferência |
| Tipo de acesso: | acesso embargado |
| Instituição associada: | Universidade Nova de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório Institucional da UNL |
| Resumo: | Healthcare plays an increasingly vital role in our daily lives. Modern Hospital Information Systems (HISs) record and store detailed medical treatment processes for all patients in the form of event logs. Leveraging these logs, process mining techniques have been widely applied to extract valuable insights, optimize medical processes, and enhance healthcare service delivery. However, existing process discovery techniques struggle to effectively handle healthcare event logs, as these processes often involve a high frequency of duplicate tasks. To address these challenges, we propose a novel duplicate task-aware process discovery technique for healthcare. More specifically, it first analyzes contextual relations among tasks to identify the sequence of duplicate tasks. Then, duplicate tasks are relabeled using a transition system derived from the input event log. Finally, a directly-follows graph is generated based on transition adjacency relations and transformed into a Petri net using the Inductive Miner. The proposed technique is fully implemented in the opensource process mining platform ProM and evaluated using a public healthcare process case with six event logs. Comparative analysis with state-of-the-art discovery techniques demonstrates that our approach accurately identifies duplicate tasks and produces high-quality process models, achieving superior replay precision and reduced complexity. |
|---|
Atividades financiadas
Não foram encontrados projetos financiados.