Sustainability | Free Full-Text | Seismic Vulnerability Assessment and Mapping of Gyeongju, South Korea using Frequency Ratio, Decision Tree, and Random Forest


The principle function of this examine was to check the prediction accuracies of varied seismic vulnerability evaluation and mapping strategies. We utilized the frequency ratio (FR), determination tree (DT), and random forest (RF) strategies to seismic knowledge for Gyeongju, South Korea. A magnitude 5.eight earthquake occurred in Gyeongju on 12 September 2016. Buildings broken through the earthquake had been used as dependent variables, and 18 sub-indicators associated to seismic vulnerability had been used as unbiased variables. Seismic knowledge had been used to assemble a mannequin for every methodology, and the fashions’ outcomes and prediction accuracies had been validated utilizing receiver working attribute (ROC) curves. The success charges of the FR, DT, and RF fashions had been 0.661, 0.899, and 1.000, and their prediction charges had been 0.655, 0.851, and 0.949, respectively. The significance of every indicator was decided, and the height floor acceleration (PGA) and distance to epicenter had been discovered to have the best affect on seismic vulnerability within the DT and RF fashions. The constructed fashions had been utilized to all buildings in Gyeongju to derive prediction values, which had been then normalized to between Zero and 1, after which divided into 5 lessons at equal intervals to create seismic vulnerability maps. An evaluation of the category distribution of constructing harm in every of the 23 administrative districts confirmed that district 15 (Wolseong) was probably the most weak space and districts 2 (Gangdong), 18 (Yangbuk), and 23 (Yangnam) had been the most secure areas.

Source link


Please enter your comment!
Please enter your name here