fix: 演示页面算法名称对齐项目、数据加噪声避免R²虚高

data/demo_equipment_costs.csv

@@ -1,27 +1,41 @@
 name,type,length_m,width_m,height_m,weight_kg,max_range_km,payload_kg,max_speed_kmh,endurance_min,tech_level,scale_level,supply_chain_level,complexity_score,actual_cost
-隼击-A,巡飞弹,1.2,1.8,0.32,18,35,4,145,55,6.4,5.8,6.2,5.9,420000
+幻影-G,巡飞弹,1.2,2.0,0.28,16,60,3,200,80,7.0,6.0,6.5,7.0,1900000
-隼击-B,巡飞弹,1.5,2.1,0.36,26,48,6,160,70,6.8,6.2,6.4,6.6,610000
+隼击-A,巡飞弹,1.2,1.49,0.29,16.21,37.98,4.25,165.47,46.82,6.22,4.82,5.57,50000,420000
-隼击-C,巡飞弹,1.8,2.5,0.42,34,65,8,175,85,7.2,6.4,6.8,7.1,830000
+隼击-B,巡飞弹,1.5,1.74,0.32,27.4,48.78,5.4,165.14,77.8,5.59,6.88,6.86,50000,610000
-侦察-100,巡飞弹,0.9,1.4,0.25,9,18,2,110,35,5.4,5.1,5.5,4.8,190000
+隼击-C,巡飞弹,1.8,2.19,0.49,32.0,55.47,6.84,196.89,88.17,8.0,6.93,6.89,50000,830000
-侦察-200,巡飞弹,1.1,1.7,0.29,14,28,3,125,48,5.9,5.4,5.7,5.3,310000
+侦察-100,巡飞弹,0.9,1.34,0.25,10.07,18.77,2.26,113.06,37.58,4.52,4.6,5.08,50000,190000
-侦察-300,巡飞弹,1.4,2.0,0.34,22,42,5,150,62,6.3,5.9,6.0,6.1,520000
+侦察-200,巡飞弹,1.1,1.54,0.25,12.88,29.37,2.85,119.16,42.98,5.41,6.25,6.0,50000,310000
-锐蛇-S,巡飞弹,1.7,2.4,0.38,30,58,7,185,76,7.5,6.7,6.9,7.4,940000
+侦察-300,巡飞弹,1.4,1.76,0.37,19.33,40.18,5.88,157.56,63.27,6.72,6.63,6.6,50000,520000
-锐蛇-M,巡飞弹,2.0,2.8,0.46,44,82,10,205,94,8.0,7.1,7.3,8.0,1360000
+锐蛇-S,巡飞弹,1.7,2.0,0.35,27.49,51.97,8.12,210.07,70.93,7.92,6.45,7.93,50000,940000
-锐蛇-L,巡飞弹,2.4,3.2,0.55,62,120,15,230,125,8.7,7.5,7.8,8.8,2100000
+锐蛇-M,巡飞弹,2.0,2.56,0.42,44.97,75.0,10.3,234.36,90.6,7.19,8.37,7.33,50000,1360000
-鹰眼-1,巡飞弹,1.3,1.9,0.31,20,40,4,155,58,6.6,5.7,6.3,6.0,470000
+锐蛇-L,巡飞弹,2.4,2.68,0.47,64.84,132.62,14.58,193.86,119.67,10.25,7.58,9.12,50000,2100000
-鹰眼-2,巡飞弹,1.6,2.2,0.37,29,57,7,172,78,7.1,6.1,6.6,6.9,760000
+鹰眼-1,巡飞弹,1.3,1.57,0.33,21.31,40.53,3.66,162.87,49.89,6.45,5.61,7.33,50000,470000
-鹰眼-3,巡飞弹,2.1,2.9,0.49,51,95,12,215,105,8.2,7.0,7.2,8.1,1580000
+鹰眼-2,巡飞弹,1.6,2.01,0.37,25.65,65.47,7.93,159.52,81.9,7.38,5.34,7.22,50000,760000
-雷霆-122,火箭炮,6.9,2.4,2.8,13500,22,480,72,0,5.8,6.6,6.0,5.5,980000
+鹰眼-3,巡飞弹,2.1,3.19,0.5,41.83,88.99,9.92,248.21,119.32,9.18,6.51,6.05,50000,1580000
-雷霆-160,火箭炮,7.6,2.6,3.0,16800,40,760,68,0,6.4,6.9,6.3,6.1,1450000
+飞燕-1,巡飞弹,1.0,1.74,0.23,10.94,18.59,2.19,131.48,34.65,5.55,5.29,5.13,50000,250000
-雷霆-220,火箭炮,8.3,2.8,3.2,21500,70,1200,65,0,7.0,7.1,6.8,7.0,2380000
+飞燕-2,巡飞弹,1.3,1.75,0.3,24.34,48.72,5.35,153.82,80.12,6.85,5.77,6.14,50000,580000
-雷霆-300,火箭炮,9.8,3.0,3.4,28500,120,1850,62,0,7.8,7.4,7.2,8.0,4200000
+飞燕-3,巡飞弹,1.9,2.34,0.41,39.21,65.0,8.09,164.88,92.15,6.86,7.79,7.91,50000,1080000
-山猫-95,火箭炮,6.2,2.3,2.7,11800,18,360,78,0,5.4,6.0,5.7,5.0,740000
+游隼-X,巡飞弹,2.2,2.72,0.55,50.24,93.7,15.04,230.75,110.9,9.37,8.22,6.75,50000,1800000
-山猫-120,火箭炮,6.7,2.4,2.8,13000,30,520,75,0,5.9,6.2,6.0,5.6,1050000
+堡垒-K,火箭炮,8.0,2.6,3.0,22000,50,800,70,0,6.0,6.5,6.0,6.0,5200000
-山猫-200,火箭炮,7.9,2.7,3.1,19800,60,980,70,0,6.8,6.8,6.5,6.7,1980000
+弩机-T,火箭炮,7.5,2.5,2.9,16000,100,1400,74,0,8.0,7.0,7.5,8.5,1950000
-山猫-300,火箭炮,9.3,2.9,3.3,26000,105,1600,66,0,7.6,7.2,7.0,7.8,3560000
+雷霆-122,火箭炮,6.9,2.34,2.72,13339.74,23.81,509.96,84.55,0,4.96,6.37,5.65,50000,980000
-弓兵-L,火箭炮,8.8,2.9,3.2,23500,85,1350,69,0,7.2,7.0,6.9,7.3,2860000
+雷霆-160,火箭炮,7.6,2.36,2.67,16489.21,38.88,699.41,61.88,0,7.38,6.76,7.12,50000,1450000
-弓兵-X,火箭炮,10.2,3.1,3.6,31000,150,2100,60,0,8.4,7.8,7.6,8.7,5400000
+雷霆-220,火箭炮,8.3,2.35,3.78,24100.85,81.82,1384.19,73.16,0,6.16,7.06,6.1,50000,2380000
-长矛-1,火箭炮,7.1,2.5,2.9,14200,28,560,73,0,6.1,6.4,6.1,5.8,1180000
+雷霆-300,火箭炮,9.8,2.52,3.25,33479.27,109.86,2039.19,61.0,0,7.58,8.62,8.48,50000,4200000
-长矛-2,火箭炮,8.1,2.7,3.1,20500,75,1120,68,0,7.1,6.9,6.7,7.1,2420000
+山猫-95,火箭炮,6.2,2.48,2.36,10936.41,21.04,370.26,79.18,0,5.88,5.04,5.87,50000,740000
-长矛-3,火箭炮,9.6,3.0,3.5,29200,130,1900,63,0,8.1,7.5,7.4,8.3,4650000
+山猫-120,火箭炮,6.7,2.7,2.45,15156.45,25.47,461.19,77.57,0,6.27,5.61,5.18,50000,1050000
-擎天-M,火箭炮,10.8,3.2,3.8,34800,180,2450,58,0,8.9,8.0,7.9,9.2,6900000
+山猫-200,火箭炮,7.9,2.45,3.21,20650.95,58.26,1009.52,70.57,0,7.86,6.08,7.01,50000,1980000
+山猫-300,火箭炮,9.3,2.79,3.5,24127.97,98.05,1745.07,55.84,0,7.49,8.49,8.25,50000,3560000
+弓兵-L,火箭炮,8.8,2.6,2.93,27165.37,96.65,1534.33,65.76,0,6.31,7.84,7.41,50000,2860000
+弓兵-X,火箭炮,10.2,3.64,3.8,25507.31,167.12,1948.33,63.53,0,9.73,6.77,6.55,50000,5400000
+长矛-1,火箭炮,7.1,2.55,2.66,14735.89,30.19,500.25,76.53,0,5.58,6.37,6.99,50000,1180000
+长矛-2,火箭炮,8.1,2.3,3.01,18851.9,61.6,1229.32,71.36,0,6.49,7.5,6.82,50000,2420000
+长矛-3,火箭炮,9.6,2.47,2.96,33227.21,148.9,1931.18,70.59,0,8.34,6.55,6.41,50000,4650000
+擎天-M,火箭炮,10.8,3.66,4.21,39318.88,205.85,2194.29,52.77,0,7.63,8.81,8.99,50000,6900000
+铁拳-80,火箭炮,6.5,2.5,2.28,14434.46,27.15,492.0,84.5,0,6.37,4.81,6.29,50000,860000
+铁拳-140,火箭炮,7.3,2.89,3.22,17531.48,61.15,806.12,78.35,0,5.33,7.37,7.0,50000,1680000
+铁拳-250,火箭炮,9.1,3.23,3.25,23060.74,105.1,1370.94,53.03,0,6.58,6.66,8.03,50000,3200000
+风暴-1,火箭炮,10.5,2.85,3.89,31326.24,193.58,2279.33,66.01,0,7.41,9.0,6.63,50000,6200000
+火焰-S,火箭炮,7.8,2.38,2.66,18357.13,51.95,699.69,74.75,0,6.63,6.42,6.58,50000,1520000
+火焰-M,火箭炮,8.6,3.01,2.71,26293.12,89.22,1348.74,72.84,0,7.75,6.85,5.92,50000,2680000
+火焰-L,火箭炮,9.9,2.82,3.27,34096.15,151.08,1809.83,56.81,0,7.93,7.33,6.86,50000,4880000

frontend/dist/index.html

@@ -6,8 +6,8 @@
     <meta name="viewport" content="width=device-width,initial-scale=1.0">
     <link rel="icon" href="/favicon.ico">
     <title>装备成本估算系统</title>
-    <script type="module" crossorigin src="/assets/index-Bh8QJqs0.js"></script>
+    <script type="module" crossorigin src="/assets/index-BclX5sLE.js"></script>
-    <link rel="stylesheet" crossorigin href="/assets/index-B09nwhwe.css">
+    <link rel="stylesheet" crossorigin href="/assets/index-D3ds-r_i.css">
   </head>
   <body>
     <noscript>

frontend/src/views/AlgorithmDemoPage.vue

@@ -178,7 +178,7 @@ import { getDemoAlgorithms, getDemoDataset, runAlgorithmDemo } from '@/api'
 
 const algorithms = ref([])
 const dataset = ref({})
-const selectedAlgorithms = ref(['linear', 'ridge', 'random_forest', 'gradient_boosting'])
+const selectedAlgorithms = ref(['linear', 'ridge', 'pls', 'random_forest', 'gbm', 'xgboost', 'lightgbm'])
 const activeAlgorithm = ref('random_forest')
 const result = ref(null)
 const loading = ref(false)
@@ -202,7 +202,7 @@ const activeMetric = computed(() => {
 })
 
 const selectRecommended = () => {
-  selectedAlgorithms.value = ['linear', 'ridge', 'random_forest', 'gradient_boosting']
+  selectedAlgorithms.value = ['linear', 'ridge', 'pls', 'random_forest', 'gbm', 'xgboost', 'lightgbm']
 }
 
 const loadInitialData = async () => {

src/demo_service.py

@@ -10,6 +10,7 @@ from sklearn.ensemble import GradientBoostingRegressor, RandomForestRegressor
 from sklearn.linear_model import LinearRegression, Ridge
 from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score
 from sklearn.model_selection import train_test_split
+from sklearn.cross_decomposition import PLSRegression
 from sklearn.neighbors import KNeighborsRegressor
 from sklearn.pipeline import Pipeline
 from sklearn.preprocessing import StandardScaler
@@ -172,49 +173,57 @@ class DemoModelService:
                 "linear",
                 "线性回归",
                 "Linear Regression",
-                "线性模型",
+                "对比基准",
-                "快速建立基准模型，用于展示参数与成本之间的线性关系。",
+                "最简单的线性基准，用于对比其他算法的提升程度。",
                 Pipeline([("scaler", StandardScaler()), ("model", LinearRegression())]),
             ),
             "ridge": AlgorithmDefinition(
                 "ridge",
                 "岭回归",
                 "Ridge Regression",
-                "线性模型",
+                "对比基准",
-                "带正则化的线性模型，适合特征存在相关性的场景。",
+                "带 L2 正则化的线性模型，适合处理共线性特征。",
                 Pipeline([("scaler", StandardScaler()), ("model", Ridge(alpha=1.0))]),
             ),
+            "pls": AlgorithmDefinition(
+                "pls",
+                "PLS回归",
+                "PLS Regression",
+                "线性模型",
+                "偏最小二乘回归，将特征投射到低维正交空间，适合小样本/共线性数据。",
+                Pipeline([("scaler", StandardScaler()), ("model", PLSRegression(n_components=4))]),
+            ),
             "random_forest": AlgorithmDefinition(
                 "random_forest",
                 "随机森林",
                 "Random Forest",
-                "树模型集成",
+                "集成学习",
-                "通过多棵决策树集成预测，能够捕捉非线性特征影响。",
+                "多棵决策树集成，能捕捉非线性关系，不易过拟合。",
-                RandomForestRegressor(n_estimators=160, max_depth=6, random_state=42),
+                RandomForestRegressor(n_estimators=100, max_depth=5, random_state=42),
             ),
-            "gradient_boosting": AlgorithmDefinition(
+            "gbm": AlgorithmDefinition(
-                "gradient_boosting",
+                "gbm",
-                "梯度提升树",
+                "GBM",
                 "Gradient Boosting",
-                "树模型集成",
+                "集成学习",
-                "逐步修正误差的提升模型，常用于表格数据回归任务。",
+                "梯度提升树，逐步修正残差，表格数据常用方法。",
-                GradientBoostingRegressor(n_estimators=120, learning_rate=0.06, max_depth=3, random_state=42),
+                GradientBoostingRegressor(n_estimators=100, learning_rate=0.08, max_depth=3, random_state=42),
             ),
             "svr": AlgorithmDefinition(
                 "svr",
                 "支持向量回归",
-                "Support Vector Regression",
+                "SVR",
-                "核方法",
+                "对比基准",
-                "使用核函数拟合平滑回归关系，适合展示不同算法偏好。",
+                "核方法回归，用于算法多样性对比。",
                 Pipeline([("scaler", StandardScaler()), ("model", SVR(C=500000, epsilon=50000))]),
             ),
             "knn": AlgorithmDefinition(
                 "knn",
-                "近邻回归",
+                "K近邻回归",
                 "KNN Regression",
-                "实例学习",
+                "对比基准",
-                "基于相似样本进行预测，便于解释局部相似性。",
+                "基于相似样本预测，便于解释局部规律。",
-                Pipeline([("scaler", StandardScaler()), ("model", KNeighborsRegressor(n_neighbors=4))]),
+                Pipeline([("scaler", StandardScaler()), ("model", KNeighborsRegressor(n_neighbors=3))]),
             ),
         }
         warnings = []
@@ -226,12 +235,12 @@ class DemoModelService:
                 "xgboost",
                 "XGBoost",
                 "XGBoost",
-                "提升模型",
+                "集成学习",
-                "面向表格数据的高性能梯度提升实现。",
+                "高性能梯度提升实现，表格数据常用。",
                 XGBRegressor(
-                    n_estimators=120,
+                    n_estimators=80,
                     max_depth=3,
-                    learning_rate=0.05,
+                    learning_rate=0.08,
                     subsample=0.9,
                     colsample_bytree=0.9,
                     random_state=42,
@@ -248,12 +257,13 @@ class DemoModelService:
                 "lightgbm",
                 "LightGBM",
                 "LightGBM",
-                "提升模型",
+                "集成学习",
-                "基于直方图优化的快速梯度提升模型。",
+                "直方图优化的快速梯度提升，大数据下效率高。",
                 LGBMRegressor(
-                    n_estimators=120,
+                    n_estimators=80,
-                    learning_rate=0.05,
+                    learning_rate=0.08,
                     max_depth=4,
+                    min_child_samples=3,
                     random_state=42,
                     verbose=-1,
                 ),