# coding=utf-8 # kNN-约会网站约友分类 from numpy import * import matplotlib.pyplot as plt import matplotlib.font_manager as font import operator # 【1】获取数据 def init_data(): # 打开训练集文件 f = open(r"datingTestSet2.txt", "r") rows = f.readlines() lines_number = len(rows) return_mat = zeros((lines_number, 3)) # lines_number行 3列 class_label_vec = [] index = 0 for row in [value.split(" ") for value in rows]: return_mat[index, :] = row[0:3] # 取row前三列 class_label_vec.append(int(row[-1])) # row[-1]取列表最后一列数据 index += 1 # 关闭打开的文件 f.close() return return_mat, class_label_vec # 【2】特征缩放 X:=[X-mean(X)]/std(X) || X:=[X-min(X)]/max(X)-min(X) ; def feature_scaling(data_set): # 特征缩放参数 max_value = data_set.max(0) min_value = data_set.min(0) # avg_value = (min_value + max_value)/2 diff_value = max_value - min_value norm_data_set = zeros(shape(data_set)) # 初始化与data_set结构一样的零array # print(norm_data_set) m = data_set.shape[0] norm_data_set = data_set - tile(min_value, (m, 1)) # avg_value norm_data_set = norm_data_set/tile(diff_value, (m, 1)) return norm_data_set, diff_value, min_value # 【3】kNN实现 input_set:输入集 data_set：训练集 def classify0(input_set, data_set, labels, k): data_set_size = data_set.shape[0] # 计算距离tile 重复以input_set生成跟data_set一样行数的mat diff_mat = tile(input_set, (data_set_size, 1)) - data_set sq_diff_mat = diff_mat ** 2 sq_distances = sq_diff_mat.sum(axis=1) distances = sq_distances ** 0.5 # 按照距离递增排序 sorted_dist_indicies = distances.argsort() # argsort返回从小到大排序的索引值 class_count = {} # 初始化一个空字典 # 选取距离最小的k个点 for i in range(k): vote_ilabel = labels[sorted_dist_indicies[i]] # 确认前k个点所在类别的出现概率,统计几个类别出现次数 class_count[vote_ilabel] = class_count.get(vote_ilabel, 0) + 1 # 返回前k个点出现频率最高的类别作为预测分类 sorted_class_count = sorted(class_count.items(), key=operator.itemgetter(1), reverse=True) return sorted_class_count[0][0] # 【4】测试kNN def dating_class_test(): # 测试样本比例 ho_ratio = 0.1 dating_data_mat, dating_labels = init_data() norm_mat, diff_dt, min_value = feature_scaling(dating_data_mat) m = norm_mat.shape[0] num_test_vecs = int(m * ho_ratio) # 测试样本的数量 error_count = 0.0 for i in range(num_test_vecs): # 测试样本和训练样本 classifier_result = classify0(norm_mat[i, :], norm_mat[num_test_vecs:m, :], dating_labels[num_test_vecs:m], 4) print("the classifier came back with:%d , the real answer is:%d" % (classifier_result, dating_labels[i])) if classifier_result != dating_labels[i]: error_count += 1.0 right_ratio = 1-error_count/float(num_test_vecs) print("the total right rate is :%f %%" % (right_ratio*100)) # 【5】样本数据绘图 def make_plot(): # 获取数据 x, y = init_data() # 特征缩放 norm_mat, diff_dt, min_value = feature_scaling(x) fig = plt.figure() ax = fig.add_subplot(111) # 画布分割一行一列数据在第一块 # 设置字体 simsun = font.FontProperties(fname='C:WindowsFontssimsun.ttc') # ax.scatter(x[:, 1], x[:, 2], 15.0*array(y), 15.0*array(y)) # 取2 3列绘图 # plt.xlabel("玩视频耗时百分比", fontproperties=simsun) # plt.ylabel("周消耗冰激凌公升数", fontproperties=simsun) ax.scatter(norm_mat[:, 0], norm_mat[:, 1], 15.0*array(y), 15.0*array(y)) # 取1 2列绘图 plt.xlabel("飞行常客里程数", fontproperties=simsun) plt.ylabel("玩视频耗时百分比", fontproperties=simsun) plt.show() # 预测函数 def classify_main(): result_list = ['not at all', 'in small doses', 'in large doses'] # 输入 ff_miles = float(input("frequent flier miles earned per year?")) percent_tats = float(input("percentage of time spent playing video games?")) ice_cream = float(input("liters of ice cream consumed per year?")) # 获取数据 dating_data_mat, dating_labels = init_data() # 特征缩放 norm_mat, diff_dt, min_value = feature_scaling(dating_data_mat) in_arr = array([ff_miles, percent_tats, ice_cream]) # 计算距离 classifier_result = classify0((in_arr-min_value)/diff_dt, norm_mat, dating_labels, 3) print("You will probably like this person:", result_list[classifier_result-1]) # 主方法 if __name__ == "__main__": # 绘图 make_plot() # 测试kNN脚本 # dating_class_test() # 预测函数 classify_main()