<![CDATA[The visa processing workflow for cruise ship crews frequently encounters delays caused by the high volume of applications and the limited availability of embassy appointment quotas, where estimation processes are still primarily conducted manually. This study aims to analyze the visa processing workflow, identify process bottlenecks, and predict processing duration as well as future surges in visa application volumes. A hybrid approach integrating Process Mining and Machine Learning is proposed in this study using historical cruise ship crew visa application data from 2022 to 2025. Process Mining was utilized to map the actual workflow and identify the primary bottleneck occurring during the visa appointment waiting stage. Furthermore, K-Means Clustering was applied to identify visa application patterns based on submission timing characteristics and process complexity, resulting in three clusters: urgent applications, early applications, and complex visa types. The final stage involved the implementation of a Linear Regression model to predict visa processing duration and application volume surges. The evaluation results demonstrate satisfactory model performance, where the visa processing duration prediction model achieved an MAE of 1.96, RMSE of 2.58, and an R-squared value of 0.65. Meanwhile, the application volume surge prediction model achieved an MAE of 5.65, RMSE of 6.74, and an R-squared value of 0.80. The findings indicate that the integration of Process Mining and Machine Learning has the potential to improve visa management efficiency through data-driven recommendations, such as scheduling document collection at least 60 days prior to departure.]]>
Copyrights © 2026
