IoT (Internet of Things) has the capability of capturing dynamic context from the physical world into the digital world. Context-aware BPM (Business Process Management) should integrate IoT as a key perspective of dynamic context of a business process and to enhance the decision making in a business process. IoT is often used to automate the process execution or integrated in the process model as resources of smart devices and additional concepts. In this way, IoT data is directly used without processing or reasoning with other contextual data to obtain higher-order contextual knowledge, which impairs its potential capability. The context layer and the decision layer are still missing while integrating IoT in BPM to obtain context-awareness. Decisions are still considered within context-aware BPM in a traditional way. This paper provides a separate concern of decisions from the process flow. We propose that the context-aware BPM ecosystem consists of four components which are: context-aware process models, context models, decision models and contextaware process execution. A framework is proposed to connect the IoT infrastructure to the context-aware BPM ecosystem using IoT-integrated ontologies and IoT-enhanced decision models, which enables the capabilities of IoT to make business processes and the decision making involved aware of the dynamic context.

The fraud detection of cargo theft has been a serious issue in ports for a long time. Traditional research in detecting theft risk is expert- and survey-based, which is not optimal for proactive prediction. As we move into a pervasive and ubiquitous paradigm, the implications of external environment and system behavior are continuously captured as multi-source data. Therefore, we propose a novel data-driven approach for formulating predictive models for detecting bulk cargo theft in ports. More specifically, we apply various feature-ranking methods and classification algorithms for selecting an effective feature set of relevant risk elements. Then, implicit Bayesian networks are derived with the features to graphically present the relationship with the risk elements of fraud. Thus, various binary classifiers are compared to derive a suitable predictive model, and Bayesian network performs best overall. The resulting Bayesian networks are then comparatively analyzed based on the outcomes of model validation and testing, as well as essential domain knowledge. The experimental results show that predictive models are effective, with both accuracy and recall values greater than 0.8. These predictive models are not only useful for understanding the dependency between relevant risk elements, but also for supporting the strategy optimization of risk management.

No abstract