Identification and hierarchical structure of cause factors for fire following earthquake using data mining and interpretive structural modeling

Historic events and prior research confirm that fire following earthquake (FFE) can cause major social and economic losses in a community. FFE is influenced by a number of interacting factors. This paper identifies 27 cause factors (CFs) for FFE through data mining method and literature review. The CFs are grouped into four clusters: management, source of ignition, environmental factors, and earthquake hazard. Interpretive Structural Modeling (ISM) is used to construct the hierarchy structure of the CFs and analyze their internal relationships. As a result, a five-level ISM is built, in which, the direct, indirect, and source of CFs are identified. Subsequently, MICMAC (cross-impact matrix multiplication applied to classification) analysis is completed to partition the CFs into four quadrants (independent, linkage, autonomous, and dependent) based on their effect index and dependence index, and evaluate the degree of relationship between the CFs. The findings show that the causal influence network with 27 CFs has a strong hierarchy, with the CFs propagating unidirectionally from the bottom layer to the top layer. The CFs in the ignition category are more dependent and influenced by other categories as expected. Investing in a resilient electric network, enhancing design standard of buildings and appropriate retrofitting, and optimizing fire prevention strategies considering seasonal hazards could reduce the risk of FFE in a community. The results of this study provide insight into the interrelationships between the CFs for FFE and can be used to identify effective risk reduction strategies and improve fire safety.