Combining association rules mining with complex networks to monitor coupled risks

Due to geotechnical uncertainties, existing underground infrastructure, the construction of deep-pit foundations in dense urban areas is particularly challenging as there is a propensity for building and structural settlement to occur. Recognizing the need to proactively manage safety risks during construction, a new risk analysis approach that combines complex networks and association rules mining (ARM) is proposed. An improved Apriori algorithm is developed to unearth abnormal monitoring types. Then, complex network theory is introduced to examine the characteristics of the coupled relationships existing between different types of abnormal monitoring types. This research identifies and examines complex network measures to understand the topology of settlement networks. It is revealed that settlement networks confirm to both scale-free and small-word properties indicating that risks are not random events. This new approach of combining ARM with complex network is applied to examine deep foundation pits that are constructed for a subway project in Wuhan, China. It is demonstrated that proposed approach can successfully reveal the association rules between safety risk monitoring types and the coupling of risks. Preventative actions can therefore be undertaken in advance to mitigate against potential risks that are identified from the abnormal monitoring combinations.