风险模型

1import json2from wi_actuarial.methods.evt import rainfall_return_level34result = rainfall_return_level.rainfall_return_level(5    annual_max=[110.9, 79.7, 50.7, 99.1, 92.9, 79.5, 91.2, 63.3, 83, 94.2, 124.2, 101.9, 76.4, 73.3, 76.6, 167.2, 43.4, 73.2, 82.8, 62.2, 90, 65.8, 76.6, 66.7, 96.2, 114.5, 62.9, 71.6, 88, 60.2, 62.4, 75.2, 69.2, 72.5, 104.5, 105.1, 76.6, 61.3, 81, 77.3, 96.8, 63.5, 71.1, 70.9, 63.1, 121.7, 69.7, 99.2, 108, 129.8, 61.7, 96.3, 91.6, 122.5, 113.2, 106.6, 77, 165.9, 77.3, 63.2, 78.2, 100.1],6    AreaCode="445302",7)8print(json.dumps(result, ensure_ascii=False))

1import json2from wi_actuarial import cat34result = cat.ep_curve(5    AreaCode="445302",6    BuildingTIV=2512731,7    LocDed6All=0,8    LocLimit6All=0,9    vulnerability="civil_works",10    design_return_period=10,11)12print(json.dumps(result, ensure_ascii=False))

1# 第 3 步：合成参考费率区间。前两步的结果来自平台的运行。2aal_permille = 0.761651   # 巨灾模型 cat.ep_curve：年平均损失率（‰）3amount = 2512730.69           # 保险金额：合同金额（元）4T = 10                          # 暴雨重现期（设防标准，年）5months = 12                     # 工期（月）6season = 1            # 施工各月的常年降水占全年的比例之和7share = (0.4, 0.6)          # 汛期施工的分项占合同金额的比例（假设）8mgmt = (0.9, 1.2)           # 施工管理系数（取值待校准）9loading = 0.35                # 费用和利润占毛保费的比例（假设）1011p_hit = 1 - (1 - 1 / T) ** (months / 12)12pure = [aal_permille * season * share[k] * mgmt[k] for k in (0, 1)]13rate = [p / (1 - loading) for p in pure]          # 参考费率区间（‰）14premium = [r / 1000 * amount for r in rate]       # 参考保费区间（元）15print(f"施工期内至少遇到一次 {T} 年一遇暴雨的概率 {p_hit:.1%}")16print(f"参考费率区间 {rate[0]:.3f}‰ 到 {rate[1]:.3f}‰，参考保费 {premium[0]:,.0f} 到 {premium[1]:,.0f} 元")