TY - JOUR UR - https://www.scopus.com/inward/record.uri?eid=2-s2.0-85115848393&doi=10.1007%2fs12665-021-09964-1&partnerID=40&md5=1c87fd26cb51638caedac03a1158b8f9 A1 - Maqsoom, A. A1 - Aslam, B. A1 - Awais, M. A1 - Hassan, U. A1 - Alaloul, W.S. A1 - Musarat, M.A. A1 - Qureshi, M.I. N1 - cited By 2 ID - scholars14414 Y1 - 2021/// IS - 19 KW - Decision theory; Earthquakes; Mapping; Operations research; Regression analysis KW - Areas under the curves; Earthquake susceptibility mapping; Hybrid model; Logistics regressions; Multi-criteria evaluation; Multi-criterion evaluation; Natural disasters; Relative indices; Seismic relative index; Susceptibility mapping KW - Disasters KW - digital mapping; earthquake; hybrid; infrastructure planning; mapping method; modeling; regression KW - Abbottabad; Khyber-Pakhtunkhwa; Pakistan; Pakistan TI - Efficiency of multiple hybrid techniques for the earthquake physical susceptibility mapping: the case of Abbottabad District, Pakistan N2 - An earthquake is a natural event that causes serious intimidation to infrastructure and humansâ?? life in northern Pakistan. Environmental, physical, and social dimensions effectively add to seismic vulnerability. The current study deals with seismic susceptibility by integrating various decisive supporting methods to generate more accurate outcomes in the Abbottabad District, Pakistan. Hybrid models: fuzzy-logistic regression (fuzzy-LR) and multi-criteria evaluationâ??logistic regression (MCEâ??LR) trained at 70 by multiple criteria decision analysisâ??multi-criteria evaluation (MCDAâ??MCE) and fuzzy-multiple criteria analysis (fuzzy-MCDA) are used to build hybrid training datasets at 30. High accuracy in the MCDAâ??MCE model is observed based on the model output. Seismic susceptibility maps are generated by implementing the resulting datasets and hybrid learning models (fuzzy-LR and MCEâ??LR). Finally, the area under the curve (AUC) and frequency ratio (FR) validate the outcomes of seismic susceptibility maps. In comparison, both MCDAâ??MCE hybrid model (AUC = 0.812) and MCEâ??LR hybrid model (AUC = 0.875) indicated more precision than fuzzy-MCDA model (AUC = 0.806) and fuzzy-LR hybrid model (AUC = 0.842), respectively. The current study concludes that training datasets are the responsible factor for defining the seismic susceptibility mapping and modelling accuracy. Moreover, this study helps to specify the high susceptible locations in the urbanized environment and facilitate policymakers to implement measures in the study area for better planning in future to avoid the effects of the earthquake. © 2021, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature. SN - 18666280 VL - 80 AV - none JF - Environmental Earth Sciences PB - Springer Science and Business Media Deutschland GmbH ER -