机器学习_感知器算法学习(Python实现)

爱被打了一巴掌 2022-12-22 09:25 128阅读 0赞

感知器是一种分类算法，代码实现如下：  
数据集提取地址：  
**链接：https://pan.baidu.com/s/1JM6mhYSuc9Rvjwk2Wgwn8Q  
提取码：56rm  
复制这段内容后打开百度网盘手机App，操作更方便哦**

import numpy as np
    import pandas as pd
    import matplotlib as mpl
    import matplotlib.pyplot as plt
    import sklearn.datasets as dataset
    
    
    class Perception:
        def __init__(self, alpha, times):
            """
            初始化方法
            :param alpha:学习率
            :param times: 迭代次数
            """
            self.alpha = alpha
            self.times = times
    
        def step(self, z):
            """
            阶跃函数
            :param z:数组类型，阶跃函数的参数，可以将z值映射为1 或者 -1
            :return: value int 类型 如果 z 大于等于0 则返回1 ，如果z小于等于0则返回-1
            """
            return np.where(z >= 0, 1, -1)
    
        def fit(self, X, y):
            X = np.asarray(X)
            y = np.asarray(y)
            # 创建权重的向量，长度比初始长度多1 多出来的为截距
            self.w_ = np.zeros(1 + X.shape[1])
            self.loss = []
            # 循环指定的次数
            for i in range(self.times):
                # 感知器与逻辑回归的区别是。逻辑回归中计算所有样本计算梯度，然后更新权重，
                # 而感知器中使用单个样本依次计算梯度，更新权重
                loss = 0
                for x, target in zip(X, y):
                    # 计算预测值
                    y_hat = np.dot(x, self.w_[1:]) + self.w_[0]
                    # 此时为 1 或者 -1
                    y_hat = self.step(y_hat)
                    # 计算损失 因为 y_hat 此时为 1 或 -1，
                    loss += y_hat != target
                    # 更新权重 w(j) = w(j) + 学习率 *（真实值　－　预测值）＊ｘ（ｊ）
                    self.w_[0] += self.alpha * (target - y_hat)
                    self.w_[1:] += self.alpha * (target - y_hat) * x
                    # print(self.w_)
                self.loss.append(loss)
    
        def predict(self, X):
            return self.step(np.dot(X, self.w_[1:]) + self.w_[0])
    
    
    if __name__ == '__main__':
        data = pd.read_csv(r'iris.csv')
        data.drop_duplicates(inplace=True)
        # print(data['class'].value_counts()) 用来查看都有什么类型Iris-versicolor Iris-virginica Iris-setosa
        data['class'] = data['class'].map({ "Iris-versicolor": 0, "Iris-virginica": 1, "Iris-setosa": -1})
        data = data[data['class'] != 0]
        t1 = data[data['class'] == 1]
        t2 = data[data['class'] == -1]
        t1 = t1.sample(len(t1), random_state=0)
        t2 = t2.sample(len(t2), random_state=0)
        train_x = np.r_[t1.iloc[:40, :-1], t2.iloc[:40, :-1]]
        train_y = np.r_[t1.iloc[:40, -1], t2.iloc[:40, -1]]
        test_x = np.r_[t1.iloc[40:, :-1], t2.iloc[40:, :-1]]
        test_y = np.r_[t1.iloc[40:, -1], t2.iloc[40:, -1]]
        p = Perception(0.1, 10)
        p.fit(train_x, train_y)
        result = p.predict(test_x)
        # print(result)
        # print(test_y)
        # print(p.loss)
        # print(p.w_)
        mpl.rcParams['font.family'] = 'SimHei'
        mpl.rcParams['axes.unicode_minus'] = False
        plt.subplot(1, 2, 1)
        plt.plot(test_y, 'go', ms=15, label='真实值')
        plt.plot(result, 'rx', ms=15, label='预测值')
        plt.title('感知器二分类')
        plt.xlabel('样本序号')
        plt.ylabel('类别')
        plt.legend()
        plt.subplot(1, 2, 2)
        plt.plot(p.loss, 'ro-')
        plt.show()