PCA降维实现

import numpy as np
from sklearn.datasets import load_iris
import pandas as pd


class myPCA():
	def __init__(self, data, k):
		self.data = np.asarray(data)
		self.n = self.data.shape[1]
		self.k = k
		self.PCA()

	def PCA(self):
		"""
		算法16.1
		"""
		# 归一化
		mean = self.data.mean(axis=1).reshape(-1, 1)
		S = (1 / (self.n - 1)) * (np.power(self.data - mean, 2)).sum(axis=1).reshape(-1, 1)
		# print(S.shape,mean.shape)
		self.data = (self.data - mean) / np.sqrt(S)
		# 构造X'
		X = (1 / np.sqrt(self.n - 1)) * self.data.T
		#print(X.shape)
		_, _, v = np.linalg.svd(X)
		#print(v.shape)
		#v=v.T
		self.result = v[:, :self.k].T @ self.data

iris = load_iris()
df = pd.DataFrame(iris.data, columns=iris.feature_names)
df['label'] = iris.target
df.columns = ['sepal length', 'sepal width', 'petal length', 'petal width', 'label']  # 原数据集标签带'(cm)'
data = np.asarray(df.iloc[:, :4]).T


myPCA = myPCA(data, 3)
print(myPCA.result.shape)