Investigating service behavior variance in port logistics from a process perspective

Abstract

This paper presents an approach that explains the synergy between process mining and data mining for the investigation of the service behavior variances in the context of port logistics. The huge variances in service behaviors are identified and regrouped by the trace clustering technique applied to the operational processes. By incorporating domain information, the unsupervised process mining result is considerably improved in both accuracy and comprehensibility. Data mining techniques are then used for investigating the correlations between the variation in services and the contributing factors. The applicability of the proposed approach is demonstrated using an extensive case study carried out at an important Chinese port. © Springer International Publishing Switzerland 2014.