📝

Author: jiangyangcreate

docs/docs/机器学习/传统算法/K均值算法.md

@@ -1,5 +1,5 @@
 ---
-sidebar_position: 1
+sidebar_position: 2
 title: K均值算法
 ---
 

docs/docs/机器学习/传统算法/支持向量机.md

@@ -1,5 +1,5 @@
 ---
-sidebar_position: 7
+sidebar_position: 8
 title: 支持向量机
 ---
 

docs/docs/机器学习/传统算法/朴素贝叶斯.md

@@ -1,5 +1,5 @@
 ---
-sidebar_position: 6
+sidebar_position: 7
 title: 朴素贝叶斯
 ---
 

docs/docs/机器学习/传统算法/线性回归.md

@@ -1,5 +1,5 @@
 ---
-sidebar_position: 4
+sidebar_position: 3
 title: 线性回归
 ---
 
@@ -29,6 +29,70 @@ title: 线性回归
 
 我们预期中，理想效果应该是 0、0 好于 -4、4 好于 7、1。只有均方误差正确的反应了这一点。
 
+通过误差的大小，我们可以慢慢修正我们的参数让线性拟合更好，导数可以反应数据变化的趋势，所以我们可以求导来修改参数。
+
+```python showLineNumbers
+import numpy as np
+from matplotlib import pyplot as plt
+
+
+class Line:
+    def __init__(self, data):
+        self.w = 1
+        self.b = 0
+        self.learning_rate = 0.01
+        self.fig, (self.ax1, self.ax2) = plt.subplots(2, 1)
+        self.loss_list = []
+
+    def get_data(self, data):
+        self.X = np.array(data)[:, 0]
+        self.y = np.array(data)[:, 1]
+
+    def predict(self, x):
+        return self.w * x + self.b
+    
+    def train(self, epoch_times):
+        for epoch in range(epoch_times):
+            total_loss = 0
+            for x, y in zip(self.X, self.y):
+                y_pred = self.predict(x)
+                # Calculate gradients
+                gradient_w = -2 * x * (y - y_pred)
+                gradient_b = -2 * (y - y_pred)
+                # Update weights
+                self.w -= self.learning_rate * gradient_w
+                self.b -= self.learning_rate * gradient_b
+                # Calculate loss
+                loss = (y - y_pred) ** 2
+                total_loss += loss
+            epoch_loss = total_loss / len(self.X)
+            self.loss_list.append(epoch_loss)
+            if epoch % 10 == 0:
+                print(f"loss: {epoch_loss}")
+                self.plot()
+        plt.ioff()
+        plt.show()
+
+    def plot(self):
+        plt.ion()  # Enable interactive mode
+        self.ax2.clear()
+        self.ax1.clear()
+        x = np.linspace(0, 10, 100)
+        self.ax1.scatter(self.X, self.y, c="g")
+        self.ax1.plot(x, self.predict(x), c="b")
+        self.ax2.plot(list(range(len(self.loss_list))), self.loss_list)
+        plt.show()
+        plt.pause(0.1)
+
+if __name__ == "__main__":  
+    # Input data
+    data = [(1, 1), (1.8, 2), (2.5, 3), (4.2, 4), (5, 5), (6, 6), (7, 7)]
+    s = Line(data)
+    s.get_data(data)
+    s.train(100)
+```
+
+## 使用sklearn模块完成
 
 ```python showLineNumbers
 import numpy as np

docs/docs/机器学习/传统算法/逻辑回归.md

@@ -1,5 +1,5 @@
 ---
-sidebar_position: 3
+sidebar_position: 4
 title: 逻辑回归
 ---
 
@@ -31,6 +31,68 @@ $y = f(β0 + β1x1 + β2x2+… βnxn)$
 
 逻辑回归分类器更接近 KNN，要解决多分类问题时，常常需要针对不同类别分别建立多个模型。
 
+```python showLineNumbers
+import numpy as np
+from matplotlib import pyplot as plt
+
+class Sline:
+    def __init__(self, data):
+        self.w = 0
+        self.b = 0
+        self.learning_rate = 0.1
+        self.fig, (self.ax1, self.ax2) = plt.subplots(2, 1)
+        self.loss_list = []
+
+
+    def get_data(self, data):
+        self.X = np.array(data)[:, 0]
+        self.y = np.array(data)[:, 1]
+
+    def sigmoid(self, x):
+        return 1 / (1 + np.exp(-(self.w * x + self.b)))
+
+    def train(self, epoch_times):
+        for epoch in range(epoch_times):
+            total_loss = 0
+            for x, y in zip(self.X, self.y):
+                y_pred = self.sigmoid(x)
+                # w新 = w旧 - 学习率 * 梯度
+                grad = -2 * (y - y_pred) * (1 - y_pred) * y_pred * x
+                self.w = self.w - self.learning_rate * grad * x
+                # b新 = b旧 - 学习率 * 梯度
+                self.b = self.b - self.learning_rate * grad
+                loss = (y - y_pred) ** 2
+                total_loss += loss
+            epoch_loss = total_loss / len(self.X)
+            self.loss_list.append(epoch_loss)
+            if epoch % 10 == 0:
+                print(f"loss: {epoch_loss}")
+                self.plot()
+        plt.ioff()
+        plt.show()
+
+    def plot(self):
+        plt.ion()  # 启用交互模式
+        self.ax2.clear()
+        self.ax1.clear()
+        x = np.linspace(0, 10, 100)
+        self.ax1.scatter(self.X, self.y, c="g")
+        self.ax1.plot(x, self.sigmoid(x), c="b")
+        self.ax2.plot(list(range(len(self.loss_list))), self.loss_list)
+        plt.show()
+        plt.pause(0.1)
+
+if __name__ == "__main__":  
+    # 散点输入
+    data = [(1, 0), (1.8, 0), (2.5, 0), (4.2, 1), (5, 1), (6, 1), (7, 1)]
+    s = Sline(data)
+    s.get_data(data)
+    s.train(1000)
+
+```
+
+## 使用sklearn框架
+
 ```python showLineNumbers
 # 绘制逻辑回归的不同回归系数的sigmoid函数
 

docs/docs/机器学习/传统算法/降维算法.md

@@ -1,5 +1,5 @@
 ---
-sidebar_position: 2
+sidebar_position: 9
 title: 降维算法
 ---
 

docs/docs/机器学习/传统算法/随机森林.md

@@ -1,5 +1,5 @@
 ---
-sidebar_position: 8
+sidebar_position: 6
 title: 随机森林
 ---