Earthquakes have caused considerable economic damage and loss of lives. The ability to determine earthquake-prone areas has utmost importance to crisis management. Given the gravity of the issue, this research proposes to incorporate geographic information system-based fuzzy-analytic network process (FAHP) and fuzzy-analytic hierarchy process (FANP) methods to determine earthquake-prone areas in the southwest of Iran. In addition, this paper is to utilize the feature selection algorithm (FSA) to select the most relevant parameters influencing earthquakes and prepare seismic activity maps. The final maps obtained using the relevant data were compared with maps obtained using the entire data. For this aim, a fuzzy membership function was initially employed to generate a fuzzy map for different layers, after which the analytic network process and analytic hierarchy process methods were utilized for assigning proper weights to each layer and eventually procuring earthquake maps. The input information for the mentioned methods consisted of landscape, slope, elevation, lithology, land use, soil, road, and fault maps. According to the FSA approach, the most relevant parameters include distance to fault (DTF), lithology, slope, and landscape. The results of FANP and FAHP methods utilizing the receiver-operating characteristic curve showed that the FANP method has a higher accuracy for determining earthquake-prone areas. As the areas located in the northern parts of the region are more at risk. In addition, Finally, the results of earthquake maps obtained using solely the relevant data from FSA were compared with the maps obtained using the entire data. The results of FSA showed that the Best first and Greedy-Stepwise models have high accuracy (R = 0.99) and the most important data for determining earthquake-prone areas using these models are DTF, lithology, slope, aspect. Therefore, by saving time and money, it is possible to prepare maps of earthquake-prone areas with high accuracy.

地震造成了相当大的经济损失和生命损失。确定地震多发地区的能力对危机管理至关重要。鉴于问题的严重性，本研究建议结合基于地理信息系统的模糊分析网络过程 (FAHP) 和模糊层次分析过程 (FANP) 方法来确定伊朗西南部的地震多发区。此外，本文将利用特征选择算法（FSA）来选择影响地震最相关的参数并制作地震活动图。将使用相关数据获得的最终地图与使用整个数据获得的地图进行比较。为此，最初采用模糊隶属函数为不同层生成模糊图，之后利用网络分析法和层次分析法为每一层分配适当的权重，最终得到地震图。上述方法的输入信息包括景观、坡度、高程、岩性、土地利用、土壤、道路和断层图。根据 FSA 方法，最相关的参数包括断层距离 (DTF)、岩性、坡度和景观。FANP 和 FAHP 方法利用接收机操作特征曲线的结果表明，FANP 方法在确定地震多发区方面具有更高的精度。由于位于该地区北部的地区面临更大的风险。此外，最后，仅使用来自 FSA 的相关数据获得的地震图的结果与使用整个数据获得的地图进行了比较。FSA 的结果表明，Best first 和 Greedy-Stepwise 模型具有较高的准确度（R  = 0.99），使用这些模型确定地震多发区的最重要数据是 DTF、岩性、坡度、坡向。因此，通过节省时间和金钱，可以制作高精度的地震多发地区地图。