已知已下数据格式，要求预测 开闸时间 开闸时长*开闸孔数 目标水位
已知数据格式如下：
-
闸下水位csv文件，包含以下字段：
- station_id：测站编号
- time：时间
- water_level：水位
-
实测流量csv文件，包含以下字段：
- 测点编码：测点编号
- 监测日期：时间
- 流量：流量值
-
实测降雨csv文件，包含以下字段：
- 测点编码：测点编号
- 监测日期：时间
- 雨量：雨量值
-
降雨预报csv文件，包含以下字段：
- 经度：经度
- 维度：纬度
- 时间：时间
- 雨量：雨量值
- 所属区域：所属区域
-
水位工况csv文件，包含以下字段： 测点编码 监测日期 水位
问题： 1）模型的预测->业务上的处理 时长：7小时 目标水位 1m
2）预测开闸时长*开孔数量 分配原则
LABEL_TO_GATES = {1: 4, 2: 8, 3: 10, 4: 12, 5: 18, 6: 28}
28孔，全开
落潮60%的时间 落7-8小时(6个小时窗口期)， 当前水位->目标水位差值，建立时间区间关联性
3）多模型如何组织 模型x -> 多值
- 不开闸样本，根据历史经验操作，一天开2次，考虑当时实际水位环境
闸下水位csv示例
station_id time water_level
0 3018 2021-01-11 10:00:00 3.33 1 3018 2021-01-11 11:00:00 4.10 2 3018 2021-01-11 12:00:00 4.42 3 3018 2021-01-11 13:00:00 3.82 4 3018 2021-01-11 14:00:00 2.90 5 3018 2021-01-11 15:00:00 1.81 6 3018 2021-01-11 16:00:00 0.86 7 3018 2021-01-11 17:00:00 0.13 8 3018 2021-01-11 18:00:00 -0.33 9 3018 2021-01-11 19:00:00 -0.72
实测流量csv示例
测点编码 监测日期 流量
0 8164 2021-01-11 10:00:00 0.00 1 8164 2021-01-11 11:00:00 11.93 2 8164 2021-01-11 12:00:00 3.58 3 8164 2021-01-11 13:00:00 0.00 4 8164 2021-01-11 14:00:00 0.00 5 8164 2021-01-11 15:00:00 16.77 6 8164 2021-01-11 16:00:00 5.99 7 8164 2021-01-11 17:00:00 46.72 8 8164 2021-01-11 18:00:00 7.19 9 8164 2021-01-11 19:00:00 9.58
实测降雨csv示例
测点编码 监测日期 雨量 经度 维度 所属区域 0 5427 2021-01-11 09:10:00 0.0 120.615600 29.765100 虞南山区 1 5425 2021-01-11 09:10:00 0.0 NaN NaN NaN 2 5422 2021-01-11 09:10:00 0.0 120.741999 29.861333 虞南山区 3 1532 2021-01-11 09:10:00 0.0 120.500000 29.894700 绍兴平原 4 3642 2021-01-11 09:10:00 0.0 120.698655 30.276754 NaN 5 3228 2021-01-11 09:10:00 0.0 120.598589 29.615754 嵊州 6 3840 2021-01-11 09:10:00 0.0 120.844937 29.917734 虞北平原 7 2310 2021-01-11 09:10:00 0.0 120.927700 30.059000 虞北平原 8 1601 2021-01-11 09:10:00 0.0 121.190897 29.523922 嵊州 9 7308 2021-01-11 09:10:00 0.0 120.744015 30.232381 NaN
降雨预报csv示例 0 1 2 3 4 5 6 0 120.25 30.15 2025-04-23 14:00:00.0000002025-04-23 15:00:00.... 0.0 NaN NaN 绍兴平原 1 120.35 30.05 2025-04-23 14:00:00.0000002025-04-23 15:00:00.... 0.0 绍兴市 柯桥区 绍兴平原 2 120.35 30.15 2025-04-23 14:00:00.0000002025-04-23 15:00:00.... 0.0 绍兴市 柯桥区 绍兴平原 3 120.45 29.95 2025-04-23 14:00:00.0000002025-04-23 15:00:00.... 0.1 绍兴市 柯桥区 绍兴平原 4 120.45 30.05 2025-04-23 14:00:00.0000002025-04-23 15:00:00.... 0.0 绍兴市 柯桥区 绍兴平原 5 120.45 30.15 2025-04-23 14:00:00.0000002025-04-23 15:00:00.... 0.0 绍兴市 柯桥区 绍兴平原 6 120.55 29.35 2025-04-23 14:00:00.0000002025-04-23 15:00:00.... 0.0 NaN NaN 嵊州 7 120.55 29.45 2025-04-23 14:00:00.0000002025-04-23 15:00:00.... 0.0 绍兴市 嵊州市 嵊州 8 120.55 29.55 2025-04-23 14:00:00.0000002025-04-23 15:00:00.... 0.0 绍兴市 嵊州市 嵊州 9 120.55 29.65 2025-04-23 14:00:00.0000002025-04-23 15:00:00.... 0.0 绍兴市 嵊州市 嵊州
水位工况csv示例 测点编码 监测日期 水位 413 1529 2021-01-11 10:00:00 4.09 414 1529 2021-01-11 11:00:00 4.08 415 1529 2021-01-11 12:00:00 4.08 416 1529 2021-01-11 13:00:00 4.08 417 1529 2021-01-11 14:00:00 4.08 418 1529 2021-01-11 15:00:00 4.08 419 1529 2021-01-11 16:00:00 4.09 420 1529 2021-01-11 17:00:00 4.09 421 1529 2021-01-11 18:00:00 4.09 422 1529 2021-01-11 19:00:00 4.10
调令csv示例 SIGNTM 日期 调度信息 开闸时间 关闸时间 开闸时长 开闸孔数 目标水位 0 2025-07-30 16:00:11 2025-07-30 07月30日傍晚退潮时段开启28孔闸门，到潮前关闭全部闸门，预计涨潮时间07-31 01:40 NaN NaN NaN 28 无最低水位限制 1 2025-07-30 05:55:49 2025-07-30 07月30日上午退潮时段开启28孔闸门，到潮前关闭全部闸门，预计涨潮时间07-30 13:45 07:15 13:10 NaN 28 无最低水位限制 2 2025-07-29 18:44:19 2025-07-29 07月29日傍晚退潮时段开启16孔闸门，07-30 00:30关闭全部闸门，预计涨潮时间07... 19:20 00:30 NaN 16 无最低水位限制 3 2025-07-29 05:54:14 2025-07-29 07月29日上午退潮时段开启20孔闸门，07-29 12:30关闭全部闸门，预计涨潮时间07... 06:30 12:30 NaN 20 无最低水位限制 4 2025-07-28 16:01:16 2025-07-28 07月28日傍晚退潮时段开启20孔闸门，07-29 00:00关闭全部闸门，预计涨潮时间07... 17:40 00:00 NaN 20 无最低水位限制 5 2025-07-28 07:08:13 2025-07-28 07月28日上午退潮时段开启12孔闸门，开闸时间3小时，预计涨潮时间07-28 13:00 08:00 11:00 3.0 12 2.5 6 2025-07-27 14:00:41 2025-07-27 07月27日下午退潮时段开启20孔闸门，开闸时间5小时，预计涨潮时间07-27 23:55 17:00 22:00 5.0 20 2.00 7 2025-07-25 14:57:03 2025-07-25 07月25日下午退潮时段开启14孔闸门，开闸时间3小时，预计涨潮时间07-25 22:50 15:40 18:40 3.0 14 3.00 8 2025-07-19 07:55:02 2025-07-19 07月19日上午退潮时段开启6孔闸门，开闸时间3小时，预计涨潮时间07-19 16:00 08:30 11:30 3.0 6 3.40 9 2025-07-18 20:32:49 2025-07-18 07月18日晚上退潮时段开启14孔闸门，开闸时间3小时，预计涨潮时间07-19 03:30 21:20 00:20 3.0 14 3.0
Fitting 5 folds for each of 6 candidates, totalling 30 fits 最佳参数: {'classifier__C': 1, 'classifier__penalty': 'l2'} 最佳交叉验证准确率: 0.0745
测试集准确率: 0.1608
precision recall f1-score support
0.0 0.00 0.00 0.00 2
1.0 0.00 0.00 0.00 2
2.0 0.00 0.00 0.00 2
3.0 0.00 0.00 0.00 1
4.0 0.00 0.00 0.00 2
5.0 0.00 0.00 0.00 3
6.0 0.00 0.00 0.00 5
7.0 0.25 0.38 0.30 8
8.0 0.00 0.00 0.00 10
9.0 0.35 0.54 0.42 13
10.0 0.00 0.00 0.00 10
11.0 0.00 0.00 0.00 8
12.0 0.00 0.00 0.00 9
13.0 0.00 0.00 0.00 11
14.0 0.00 0.00 0.00 12
15.0 0.00 0.00 0.00 15
16.0 0.16 0.64 0.25 25
17.0 0.00 0.00 0.00 20
18.0 0.13 0.33 0.19 18
19.0 0.00 0.00 0.00 10
20.0 0.00 0.00 0.00 6
21.0 0.00 0.00 0.00 3
22.0 0.00 0.00 0.00 3
23.0 0.00 0.00 0.00 1
accuracy 0.16 199
macro avg 0.04 0.08 0.05 199 weighted avg 0.06 0.16 0.09 199
python -m models.kz_time.train
python -m models.kz_dura.train
python -m models.kz_num.train
python -m models.kz_level.train
python -m models.kz_time.evaluate
python -m models.kz_dura.evaluate
python -m models.kz_num.evaluate
python -m models.kz_level.evaluate
任务类型 模型名称
分类 逻辑回归 (Logistic Regression) 处理二分类问题，计算效率高，结果有概率意义 支持向量机 (SVM - 用于分类) 善于找到复杂分类边界，尤其适合高维数据和小样本情况 决策树 (Decision Tree - 用于分类) 模型直观易解释，容易过拟合
回归 线性回归 (Linear Regression) 建模变量间线性关系，简单、高效、可解释性强 岭回归 (Ridge Regression) 在线性回归基础上加入L2正则化，处理特征多重共线性，防止过拟合 Lasso回归 (Lasso Regression) 在线性回归基础上加入L1正则化，能进行特征选择，使模型更稀疏
分类与回归 支持向量机 (SVM) 通过不同核函数和损失函数，可灵活应用于分类和回归问题 决策树 (Decision Tree) 通过不同的分裂准则（如基尼系数、信息增益或均方误差）处理分类和回归任务 随机森林 (Random Forest) 集成多棵决策树，通过投票或平均进行预测，泛化能力强大，抗过拟合 梯度提升机 (XGBoost, LightGBM) 逐步修正前序模型的错误，预测精度高，在许多数据竞赛中表现出色
特征分布情况： hour_of_day day_of_week month is_weekend hour_sin hour_cos day_of_year prev_gate_count ... flow_rain_ratio is_rush_hour water_status_mean water_status_max water_status_min water_status_range water_status_slope water_status_missing count 980.000000 980.000000 980.000000 980.000000 980.000000 980.000000 980.000000 980.000000 ... 9.800000e+02 980.000000 980.000000 980.00000 980.000000 980.000000 9.800000e+02 980.000000 mean 12.974490 2.924490 5.750000 0.254082 -0.102901 -0.473695 159.624490 10.806122 ... 3.347122e+05 0.317347 4.587168 26.06501 1.970969 24.094041 1.674770e-07 0.381633 std 4.407066 1.958823 3.013716 0.435566 0.713413 0.506801 91.809323 4.488000 ... 1.691327e+06 0.465681 3.611599 20.49194 1.667278 18.951099 3.067141e-06 0.486035 min 0.000000 0.000000 1.000000 0.000000 -1.000000 -1.000000 2.000000 0.000000 ... 0.000000e+00 0.000000 0.000000 0.00000 -1.870000 0.000000 -2.941310e-05 0.000000 25% 9.000000 1.000000 3.000000 0.000000 -0.707107 -0.866025 84.000000 8.000000 ... 0.000000e+00 0.000000 0.000000 0.00000 0.000000 0.000000 0.000000e+00 0.000000 50% 13.000000 3.000000 6.000000 0.000000 -0.258819 -0.707107 155.000000 10.000000 ... 3.821974e-02 0.000000 7.211200 41.92500 2.640000 38.440000 0.000000e+00 0.000000 75% 16.000000 5.000000 8.000000 1.000000 0.707107 -0.258819 221.000000 14.000000 ... 6.370912e-01 1.000000 7.490090 42.40250 3.617500 39.052500 7.430340e-07 1.000000 max 23.000000 6.000000 12.000000 1.000000 1.000000 1.000000 365.000000 28.000000 ... 2.514212e+07 1.000000 8.355417 42.81000 4.000000 44.230000 2.520487e-05 1.000000
[8 rows x 97 columns]
=== 分析目标变量: 开闸时间(小时) ===
特征与目标 '开闸时间(小时)' 的相关性： hour_of_day 0.339329 tide_24h_tide_rise_rate 0.117882 tide_24h_tide_fall_rate 0.105184 future_water_missing 0.095228 future_tide_tide_slope 0.091744 tide_24h_tide_cycle_count 0.089262 future_tide_tide_range 0.080782 tide_12h_tide_fall_rate 0.064833 flow_rain_ratio 0.063483 future_tide_tide_max 0.059073 dtype: float64 随机森林模型性能 (开闸时间(小时)): R2=0.5474, MSE=10.7882
特征重要性排序 (开闸时间(小时)): feature importance 0 hour_of_day 0.391009 5 hour_cos 0.160437 4 hour_sin 0.073048 89 flow_rain_ratio 0.050522 9 prev_op_hour 0.042500 12 ops_week_total_duration 0.029763 38 future_tide_tide_slope 0.021621 6 day_of_year 0.019505 17 tide_24h_tide_slope 0.016856 1 day_of_week 0.015986
=== 分析目标变量: 开闸时长 ===
特征与目标 '开闸时长' 的相关性： ops_week_total_duration 0.481209 water_status_missing 0.380520 flow_missing 0.379216 water_missing 0.377869 ops_week_count 0.375852 tide_24h_tide_min 0.375542 prev_gate_count 0.365668 tide_12h_tide_min 0.352008 ops_week_avg_gates 0.288154 prev_duration 0.274795 dtype: float64 随机森林模型性能 (开闸时长): R2=0.3937, MSE=0.4879
特征重要性排序 (开闸时长): feature importance 93 water_status_min 0.151503 12 ops_week_total_duration 0.148799 8 prev_duration 0.079649 11 ops_week_avg_gates 0.053617 6 day_of_year 0.051068 9 prev_op_hour 0.047107 7 prev_gate_count 0.043042 1 day_of_week 0.027432 5 hour_cos 0.026518 0 hour_of_day 0.022259
=== 分析目标变量: 开闸孔数 ===
特征与目标 '开闸孔数' 的相关性： prev_gate_count 0.458496 rain_actual_avg 0.386750 rain_actual_total 0.386750 ops_week_count 0.375011 ops_week_total_duration 0.374464 rain_actual_虞北平原_sum 0.370383 rain_actual_嵊州_sum 0.366215 rain_actual_绍兴平原_sum 0.361241 rain_actual_虞南山区_sum 0.360504 ops_week_avg_gates 0.355064 dtype: float64 随机森林模型性能 (开闸孔数): R2=0.3413, MSE=2.5584
特征重要性排序 (开闸孔数): feature importance 12 ops_week_total_duration 0.141965 82 rain_actual_total 0.069843 7 prev_gate_count 0.062412 84 rain_actual_avg 0.053470 93 water_status_min 0.051714 6 day_of_year 0.050228 11 ops_week_avg_gates 0.041803 9 prev_op_hour 0.040433 39 future_tide_tide_r_squared 0.025213 1 day_of_week 0.023776
=== 分析目标变量: 目标水位 ===
特征与目标 '目标水位' 的相关性： tide_24h_tide_cycle_count 0.384441 water_status_min 0.380125 tide_24h_tide_rise_rate 0.376914 tide_24h_tide_fall_rate 0.376012 future_tide_tide_r_squared 0.340304 water_status_mean 0.270662 water_status_max 0.259960 tide_12h_tide_cycle_count 0.254384 water_status_range 0.247654 tide_24h_tide_range 0.240236 dtype: float64 随机森林模型性能 (目标水位): R2=0.2738, MSE=0.5279
特征重要性排序 (目标水位): feature importance 93 water_status_min 0.199692 7 prev_gate_count 0.196222 12 ops_week_total_duration 0.074290 6 day_of_year 0.072297 11 ops_week_avg_gates 0.046322 9 prev_op_hour 0.043468 8 prev_duration 0.042399 1 day_of_week 0.020301 5 hour_cos 0.019575 10 ops_week_count 0.019191
=== 分析目标变量: 时长孔数乘积 ===
特征与目标 '时长孔数乘积' 的相关性：
prev_gate_count 0.479105
ops_week_total_duration 0.473047
ops_week_count 0.413503
ops_week_avg_gates 0.350063
rain_actual_avg 0.257980
rain_actual_total 0.257980
tide_24h_tide_min 0.256952
rain_actual_嵊州_sum 0.247580
tide_12h_tide_min 0.245707
rain_actual_虞南山区_sum 0.244365
dtype: float64
随机森林模型性能 (时长孔数乘积): R2=0.4115, MSE=66.2193
特征重要性排序 (时长孔数乘积): feature importance 12 ops_week_total_duration 0.147010 7 prev_gate_count 0.130220 93 water_status_min 0.091072 11 ops_week_avg_gates 0.056557 6 day_of_year 0.056056 9 prev_op_hour 0.050736 8 prev_duration 0.048958 10 ops_week_count 0.027358 1 day_of_week 0.026619 89 flow_rain_ratio 0.025566
=== 分析目标变量: 时长孔数乘积(对数) ===
特征与目标 '时长孔数乘积(对数)' 的相关性： ops_week_total_duration 0.471087 prev_gate_count 0.465736 ops_week_count 0.425601 ops_week_avg_gates 0.387984 rain_actual_avg 0.262709 rain_actual_total 0.262709 tide_24h_tide_min 0.261862 tide_12h_tide_min 0.251698 prev_duration 0.251621 rain_actual_嵊州_sum 0.251174 dtype: float64 随机森林模型性能 (时长孔数乘积(对数)): R2=0.3627, MSE=0.2361
特征重要性排序 (时长孔数乘积(对数)): feature importance 93 water_status_min 0.216666 12 ops_week_total_duration 0.088104 7 prev_gate_count 0.085818 6 day_of_year 0.053693 11 ops_week_avg_gates 0.045369 8 prev_duration 0.044574 9 prev_op_hour 0.040881 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660 12 ops_week_total_duration 0.147010 7 prev_gate_count 0.130220 93 water_status_min 0.091072 11 ops_week_avg_gates 0.056557 6 day_of_year 0.056056 9 prev_op_hour 0.050736 8 prev_duration 0.048958 10 ops_week_count 0.027358 1 day_of_week 0.026619 89 flow_rain_ratio 0.025566
=== 分析目标变量: 时长孔数乘积(对数) ===
特征与目标 '时长孔数乘积(对数)' 的相关性： ops_week_total_duration 0.471087 prev_gate_count 0.465736 ops_week_count 0.425601 ops_week_avg_gates 0.387984 rain_actual_avg 0.262709 rain_actual_total 0.262709 tide_24h_tide_min 0.261862 tide_12h_tide_min 0.251698 prev_duration 0.251621 rain_actual_嵊州_sum 0.251174 dtype: float64 随机森林模型性能 (时长孔数乘积(对数)): R2=0.3627, MSE=0.2361
特征重要性排序 (时长孔数乘积(对数)): feature importance 93 water_status_min 0.216666 12 ops_week_total_duration 0.088104 7 prev_gate_count 0.085818 6 day_of_year 0.053693 11 ops_week_avg_gates 0.045369 8 prev_duration 0.044574 9 prev_op_hour 0.040881 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660 1 day_of_week 0.025348
12 ops_week_total_duration 0.147010 7 prev_gate_count 0.130220 93 water_status_min 0.091072 11 ops_week_avg_gates 0.056557 6 day_of_year 0.056056 9 prev_op_hour 0.050736 8 prev_duration 0.048958 10 ops_week_count 0.027358 1 day_of_week 0.026619 89 flow_rain_ratio 0.025566
=== 分析目标变量: 时长孔数乘积(对数) ===
特征与目标 '时长孔数乘积(对数)' 的相关性： ops_week_total_duration 0.471087 prev_gate_count 0.465736 ops_week_count 0.425601 ops_week_avg_gates 0.387984 rain_actual_avg 0.262709 rain_actual_total 0.262709 tide_24h_tide_min 0.261862 tide_12h_tide_min 0.251698 prev_duration 0.251621 rain_actual_嵊州_sum 0.251174 dtype: float64 随机森林模型性能 (时长孔数乘积(对数)): R2=0.3627, MSE=0.2361
特征重要性排序 (时长孔数乘积(对数)): feature importance 93 water_status_min 0.216666 12 ops_week_total_duration 0.088104 7 prev_gate_count 0.085818 6 day_of_year 0.053693 11 ops_week_avg_gates 0.045369 8 prev_duration 0.044574 9 prev_op_hour 0.040881 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660 9 prev_op_hour 0.050736 8 prev_duration 0.048958 10 ops_week_count 0.027358 1 day_of_week 0.026619 89 flow_rain_ratio 0.025566
=== 分析目标变量: 时长孔数乘积(对数) ===
特征与目标 '时长孔数乘积(对数)' 的相关性： ops_week_total_duration 0.471087 prev_gate_count 0.465736 ops_week_count 0.425601 ops_week_avg_gates 0.387984 rain_actual_avg 0.262709 rain_actual_total 0.262709 tide_24h_tide_min 0.261862 tide_12h_tide_min 0.251698 prev_duration 0.251621 rain_actual_嵊州_sum 0.251174 dtype: float64 随机森林模型性能 (时长孔数乘积(对数)): R2=0.3627, MSE=0.2361
特征重要性排序 (时长孔数乘积(对数)): feature importance 93 water_status_min 0.216666 12 ops_week_total_duration 0.088104 7 prev_gate_count 0.085818 6 day_of_year 0.053693 11 ops_week_avg_gates 0.045369 8 prev_duration 0.044574 9 prev_op_hour 0.040881 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660
=== 分析目标变量: 时长孔数乘积(对数) ===
特征与目标 '时长孔数乘积(对数)' 的相关性： ops_week_total_duration 0.471087 prev_gate_count 0.465736 ops_week_count 0.425601 ops_week_avg_gates 0.387984 rain_actual_avg 0.262709 rain_actual_total 0.262709 tide_24h_tide_min 0.261862 tide_12h_tide_min 0.251698 prev_duration 0.251621 rain_actual_嵊州_sum 0.251174 dtype: float64 随机森林模型性能 (时长孔数乘积(对数)): R2=0.3627, MSE=0.2361
特征重要性排序 (时长孔数乘积(对数)): feature importance 93 water_status_min 0.216666 12 ops_week_total_duration 0.088104 7 prev_gate_count 0.085818 6 day_of_year 0.053693 11 ops_week_avg_gates 0.045369 8 prev_duration 0.044574 9 prev_op_hour 0.040881 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660 rain_actual_total 0.262709 tide_24h_tide_min 0.261862 tide_12h_tide_min 0.251698 prev_duration 0.251621 rain_actual_嵊州_sum 0.251174 dtype: float64 随机森林模型性能 (时长孔数乘积(对数)): R2=0.3627, MSE=0.2361
特征重要性排序 (时长孔数乘积(对数)): feature importance 93 water_status_min 0.216666 12 ops_week_total_duration 0.088104 7 prev_gate_count 0.085818 6 day_of_year 0.053693 11 ops_week_avg_gates 0.045369 8 prev_duration 0.044574 9 prev_op_hour 0.040881 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660 dtype: float64 随机森林模型性能 (时长孔数乘积(对数)): R2=0.3627, MSE=0.2361
特征重要性排序 (时长孔数乘积(对数)): feature importance 93 water_status_min 0.216666 12 ops_week_total_duration 0.088104 7 prev_gate_count 0.085818 6 day_of_year 0.053693 11 ops_week_avg_gates 0.045369 8 prev_duration 0.044574 9 prev_op_hour 0.040881 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660 12 ops_week_total_duration 0.088104 7 prev_gate_count 0.085818 6 day_of_year 0.053693 11 ops_week_avg_gates 0.045369 8 prev_duration 0.044574 9 prev_op_hour 0.040881 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660 11 ops_week_avg_gates 0.045369 8 prev_duration 0.044574 9 prev_op_hour 0.040881 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660 9 prev_op_hour 0.040881 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660 89 flow_rain_ratio 0.028558 39 future_tide_tide_r_squared 0.025660 39 future_tide_tide_r_squared 0.025660 1 day_of_week 0.025348
=== 目标变量之间的关系分析 === 目标变量相关性矩阵: 开闸时间(小时) 开闸时长 开闸孔数 目标水位 时长孔数乘积 时长孔数乘积(对数) 开闸时间(小时) 1.000000 -0.051780 -0.138659 0.121547 -0.097318 -0.098613 开闸时长 -0.051780 1.000000 0.493280 -0.697410 0.842913 0.806962 开闸孔数 -0.138659 0.493280 1.000000 -0.686679 0.852021 0.860638 目标水位 0.121547 -0.697410 -0.686679 1.000000 -0.801005 -0.755715 时长孔数乘积 -0.097318 0.842913 0.852021 -0.801005 1.000000 0.914692 时长孔数乘积(对数) -0.098613 0.806962 0.860638 -0.755715 0.914692 1.000000
预测开闸与否 POST 127.0.0.1:8001/predict_bin
{ "metadata": { "model_version": "binary_1.0", "prediction_time": "2025-09-24 13:41:31", "threshold_used": 0.5 }, "prediction": { "confidence": 0.7303, "confidence_level": "medium", "explanation": "预测开闸 (概率: 0.730) - 中等置信度开闸", "features_used": 159, "prediction": 1, "probability": 0.7303, "recommendation": "Issue gate opening order" }, "status": "success" }
本防汛调度经验模型是基于多源监测数据与机器学习构建的智能决策系统。 1）融合水位、流量、降雨及潮汐预报等多维数据，通过特征工程提取97项关键指标，支撑精准分析； 2）采用多模型协同架构，分别预测开闸时间、时长、孔数和目标水位，并依据潮汐规律自动识别泄洪窗口，实现孔数分级智能分配；