Softmax regression逻辑回归解决多线性分类

import numpy as np import matplotlib.pyplot as plt import random as rd # 模型训练函数 def gradientAscent(feature_data, label_data, k, maxCycle, alpha): """ 利用梯度下降法训练Softmax模型 :param feature_data:（mat）特征 :param label_data:（mat）标签 :param k:（int）类别的个数 :param maxCycle:（int）最大迭代个数 :param alpha: （float）学习速率 :return:weights（mat）权重 """ m, n = np.shape(feature_data) weights = np.mat(np.ones((n, k))) i = 0 while i <= maxCycle: err = np.exp(feature_data * weights) if i % 1000 == 0: print("\t-----iter:", i, ",cost:", cost(err, label_data)) row_sum = -err.sum(axis=1) row_sum = row_sum.repeat(k, axis=1) err = err / row_sum for x in range(m): err[x, label_data[x, 0]] += 1 weights = weights + (alpha / m) * feature_data.T * err i += 1 return weights def cost(err, label_data): """ 计算损失函数值 :param err:（mat）,exp值 :param label_data:（）mat标签的值 :return:sum_cost/m (float):损失函数的值 """ m = np.shape(err)[0] sum_cost = 0.0 for i in range(m): if err[i, label_data[i, 0]] / np.sum(err[i, :]) > 0: sum_cost -= np.log(err[i, label_data[i, 0]] / np.sum(err[i, :])) else: sum_cost -= 0 return sum_cost / m def load_data(inputfile): """ :param inputfile:训练样本的文件名 :return:feature_data(mat)特征 label_data(mat)标签 k(int)类别的个数 """ f = open(inputfile) feature_data = [] label_data = [] for line in f.readlines(): feature_temp = [] feature_temp.append(1) lines = line.strip().split("\t") for i in range(len(lines) - 1): feature_temp.append(float(lines[i])) label_data.append(int(lines[-1])) feature_data.append(feature_temp) f.close() # print(feature_data) # feature_data = np.mat(feature_data) # print(feature_data) return np.mat(feature_data), np.mat(label_data).T, len(set(label_data)) def save_model(filename, weights): """ 保存最终的模型 :param filename:(string)文件名 :param w: （mat）SR 模型的权重 :return: """ f_w = open(filename, "w") m, n = np.shape(weights) for i in range(m): w_tmp = [] for j in range(n): w_tmp.append(str(weights[i, j])) f_w.write("\t".join(w_tmp) + "\n") f_w.close() def figurePlot(feature, label): """ 根据标签画出各个类别的点图 :param feature:（mat）特征 :param label:（mat）标签 :return: """ point_red = [] point_blue = [] point_yellow = [] point_green = [] r = np.shape(feature)[0] for index in range(r): temp = -2 # 从右到左第二个数 if label[index] == 0: point_red.append(feature[index, temp]) point_red.append(feature[index, temp + 1]) # 如果一次放两个元素feature[index, 1:3]，会被当做一个整体存放在list中 elif label[index] == 1: point_blue.append(feature[index, temp]) point_blue.append(feature[index, temp + 1]) elif label[index] == 2: point_yellow.append(feature[index, temp]) point_yellow.append(feature[index, temp + 1]) else: point_green.append(feature[index, temp]) point_green.append(feature[index, temp + 1]) # 画出图形 point_red = np.mat(point_red).reshape(-1, 2) # list转换成一维矩阵，再转换成二维矩阵 point_blue = np.mat(point_blue).reshape(-1, 2) point_green = np.mat(point_green).reshape(-1, 2) point_yellow = np.mat(point_yellow).reshape(-1, 2) plt.scatter(point_red[:, 0].tolist(), point_red[:, 1].tolist(), c='r') # scatter函数只接受list，不接受matrix plt.scatter(point_blue[:, 0].tolist(), point_blue[:, 1].tolist(), c='b') plt.scatter(point_green[:, 0].tolist(), point_green[:, 1].tolist(), c='g') plt.scatter(point_yellow[:, 0].tolist(), point_yellow[:, 1].tolist(), c='y') plt.show() # 模型测试函数 def load_weights(weights_path): """ 训练好的softmax模型 :param weights_path:（string）文件名（文件的存储位置） :return:weights（mat）将权重存储的矩阵中 m（int）权重的行数 n（int）权重的列数 """ f = open(weights_path) weights = [] for line in f.readlines(): w_tmp = [] lines = line.strip().split("\t") for x in lines: w_tmp.append(float(x)) weights.append(w_tmp) f.close() print(weights) weights = np.mat(weights) m, n = np.shape(weights) return weights, m, n def load_testData(num, m): """ 导入测试数据 :param num:（int）生成测试样本的个数 :param m:（int）样本的维数 :return:testDataSet（mat）生成测试样本 """ test_DataSet = np.mat(np.ones((num, m))) for i in range(num): # 随机生成[-3,3]之间的随机数,rd.random生成(0,1)之间的随机数 test_DataSet[i, 1] = rd.random() * 6 - 3 # 随机生成[0,15]之间的随机数 test_DataSet[i, 2] = rd.random() * 15 return test_DataSet def predict(test_data, weights): """ 利用训练好的Softmax模型对测试数据进行预测 :param test_data:（mat）测试数据的特征 :param weights:（mat）模型的权重 :return:h = argmax（axis=1）##行方向的最大值所在的列数，axis=1代表行 """ h = test_data * weights return h.argmax(axis=1) # 获得所属的类别 def save_result(filename, result): """ 保存最终的预测结果 :param filename:（string）保存结果的文件名 :param result:（mat）最终的结果 :return: """ f_result = open(filename, 'w') m = np.shape(result)[0] for i in range(m): f_result.write(str(result[i, 0]) + "\n") f_result.close()