《模式识别与智能计算》基于类中心的欧式距离分类法

算法流程

1. 选取某一类样本X
2. 计算样本类中心
3. 采用欧式距离测度计算待测样品到类中心的距离
4. 距离最小的就是待测样品的类别

算法实现

计算距离

def euclid(x_train,y_train,sample):""":function: 基于类中心的模板匹配法:param x_train:训练集 M*N M为样本个数 N为特征个数:param y_train:训练集标签 1*M:param sample: 待识别样品:return: 返回判断类别"""disMin &＃61; np.inflabel &＃61; 0#去除标签重复元素target &＃61; np.unique(y_train)for i in target:#将同一类别的数据下标集中到一起trainId &＃61;([j for j,y in enumerate(y_train) if y&＃61;&＃61;i])train &＃61; x_train[trainId,:]trainMean &＃61; np.mean(train, axis&＃61;0)dis &＃61; np.dot((sample-trainMean),(sample - trainMean).T)if(disMin>dis):disMin &＃61; dislabel &＃61; ireturn label


划分数据集

def train_test_split(x,y,ratio &＃61; 3):""":function: 对数据集划分为训练集、测试集:param x: m*n维 m表示数据个数 n表示特征个数:param y: 标签:param ratio: 产生比例 train:test &＃61; 3:1(默认比例):return: x_train y_train x_test y_test"""n_samples , n_train &＃61; x.shape[0] , int(x.shape[0]*(ratio)/(1&＃43;ratio))train_id &＃61; random.sample(range(0,n_samples),n_train)x_train &＃61; x[train_id,:]y_train &＃61; y[train_id]x_test &＃61; np.delete(x,train_id,axis &＃61; 0)y_test &＃61; np.delete(y,train_id,axis &＃61; 0)return x_train,y_train,x_test,y_test


测试

from sklearn import datasets
from Include.chapter3 import function
import numpy as np#读取数据
digits &＃61; datasets.load_digits()
x , y &＃61; digits.data,digits.target#划分数据集
x_train, y_train, x_test, y_test &＃61; function.train_test_split(x,y)
testId &＃61; np.random.randint(0, x_test.shape[0])
sample &＃61; x_test[testId, :]ans &＃61; function.euclid(x_train,y_train,sample)
print(ans&＃61;&＃61;y_test[testId])

算法结果

True