从头开始进行 PCA 时特征值没有收敛

Question

我不断收到 LinAlgError 消息，指出特征值没有收敛。当我将 x_mean 值的类型更改为 float 时出现此错误，但是当该值未设置为 float 时，我会收到一条错误消息，指出“float”对象没有属性“shape”。我的代码还有其他问题吗？我是否应该将 x_mean 值更改为不同的值？

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

#Read data file
data = pd.read_csv(r"Data.csv", names = ['length',
                                                                    'width',
                                                                    'petal length',
                                                                    'petal width',
                                                                    'target'])
def iris_PCA(X, num_dimensions):
    #calculate the mean for the values
    x_mean = data_iris - np.mean(data_iris, axis = 0)

    #calculate the covariance matrix
    cov_x = np.cov(x_mean.astype(float), rowvar = False)

    #calculate the eigenvalues and vectors based on the covariance found
    e_values = np.linalg.eigh(cov_x)
    e_vectors = np.linalg.eigh(cov_x)

    #sorting the eigenvectors in descending order
    sorted_index = np.argsor(e_values)[::-1]
    sorted_e_values = e_values[sorted_index]
    sorted_e_vectors = e_values[sorted_index]
    
    e_subset = sorted_e_vectors[:,0:num_dimensions]
    
    #perform dot product to reduce data to lowered dimensions
    reduced_x = np.dot(e_subset.transpose(), x_mean.transpose()).transpose()

    return reduced_x

x = data_iris.iloc[:,0:4]
target = data_iris.iloc[:,4]

#displaying the results in the 3rd dimension
reduced_x3 = iris_PCA(x,3)

PCA_df =  pd.DataFrame(reduced_x3, columns = ['PCA1', 'PCA2', 'PCA3'])

PCA_df = pd.concat([PCA_df, pd.DataFrame(target)],axis = 1)

plt.figure(figsize = (6,6))