写文章交作业
这次训练营的讲的是knn
我先找下文章回忆下
KNN算法
KNN实现“手写识别”
GridSearchCV和交叉熵
机器学习的回归算法
sklearn模型的训练(上）
近邻算法分类
[机器学习认识聚类（KMeans算法）]
这些都是我写的吗？？？怎么没啥印象了？？？？？
确定k值
from sklearn.cluster import KMeanswcss = for k in range(1,15): kmeans = KMeans(n_clusters=k) kmeans.fit(x) wcss.append(kmeans.inertia_)plt.plot(range(1,15),wcss)plt.xlabel("k values")plt.ylabel("WCSS")plt.show 作者：毛利学python 来源：CSDN 原文：https://blog.csdn.net/weixin_44510615/article/details/92021326 版权声明：本文为博主原创文章，转载请附上博文链接！
数据采用的是经典的iris数据，是三分类问题
# 读取相应的库from sklearn import datasetsfrom sklearn.model_selection import train_test_splitfrom sklearn.neighbors import KNeighborsClassifierimport numpy as np# 读取数据 X, yiris = datasets.load_irisX = iris.datay = iris.targetprint (X, y)# 把数据分成训练数据和测试数据X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=2003)# 构建KNN模型， K值为3、 并做训练clf = KNeighborsClassifier(n_neighbors=3)clf.fit(X_train, y_train)# 计算准确率from sklearn.metrics import accuracy_scorecorrect = np.count_nonzero((clf.predict(X_test)==y_test)==True)# 35#accuracy_score(y_test, clf.predict(X_test))print ("Accuracy is: %.3f" %(correct/len(X_test))) # 35/38
np.count_nonzero是来数组的0的总数
自己手下一个knn
计算两个样本的欧式距离得到最大k个的距离再投票计数
from sklearn import datasetsfrom collections import Counter # 为了做投票from sklearn.model_selection import train_test_splitimport numpy as np# 导入iris数据iris = datasets.load_irisX = iris.datay = iris.targetX_train, X_test, y_train, y_test = train_test_split(X, y, random_state=2003)def euc_dis(instance1, instance2): """ 计算两个样本instance1和instance2之间的欧式距离 instance1: 第一个样本， array型 instance2: 第二个样本， array型 """ # TODO dist = np.sqrt(sum((instance1 – instance2)**2)) return dist def knn_classify(X, y, testInstance, k): """ 给定一个测试数据testInstance, 通过KNN算法来预测它的标签。 X: 训练数据的特征 y: 训练数据的标签 testInstance: 测试数据，这里假定一个测试数据 array型 k: 选择多少个neighbors? """ # TODO 返回testInstance的预测标签 = {0,1,2} distances = [euc_dis(x, testInstance) for x in X]) # 排序 kneighbors = np.argsort(distances)[:k] # count是一个字典 count = Counter(y[kneighbors]) # count.most_common[0][0])是票数最多的 return count.most_common[0][0]
x = np.array([3, 1, 2]) np.argsort(x)array([1, 2, 0])
Counter不会的看
python标准库 collections
# 预测结果。 predictions = [knn_classify(X_train, y_train, data, 3) for data in X_test]correct = np.count_nonzero((predictions==y_test)==True)#accuracy_score(y_test, clf.predict(X_test))print ("Accuracy is: %.3f" %(correct/len(X_test)))
KNN的决策边界
import matplotlib.pyplot as pltimport numpy as npfrom itertools import product# knn分类from sklearn.neighbors import KNeighborsClassifier# 生成一些随机样本n_points = 100# 多元高斯分布X1 = np.random.multivariate_normal([1,50], [[1,0],[0,10]], n_points)X2 = np.random.multivariate_normal([2,50], [[1,0],[0,10]], n_points)X = np.concatenate([X1,X2])y = np.array([0]*n_points + [1]*n_points)print (X.shape, y.shape) # (200, 2) (200,)
# KNN模型的训练过程clfs = neighbors = [1,3,5,9,11,13,15,17,19]for i in range(len(neighbors)): clfs.append(KNeighborsClassifier(n_neighbors=neighbors[i]).fit(X,y))
# 可视化结果x_min, x_max = X[:, 0].min – 1, X[:, 0].max + 1y_min, y_max = X[:, 1].min – 1, X[:, 1].max + 1xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.1), np.arange(y_min, y_max, 0.1))f, axarr = plt.subplots(3,3, sharex='col', sharey='row', figsize=(15, 12))for idx, clf, tt in zip(product([0, 1, 2], [0, 1, 2]), clfs, ['KNN (k=%d)'%k for k in neighbors]): Z = clf.predict(np.c_[xx.ravel, yy.ravel]) Z = Z.reshape(xx.shape) axarr[idx[0], idx[1]].contourf(xx, yy, Z, alpha=0.4) axarr[idx[0], idx[1]].scatter(X[:, 0], X[:, 1], c=y, s=20, edgecolor='k') axarr[idx[0], idx[1]].set_title(tt)
