pca百分比取多少比较好_数据处理之PCA

推荐好文PCA的数学原理

本文将会用Python来实现PCA，帮助更好的理解

1. 获取数据

我们用的数据是150个鸢尾花，然后通过4个维度刻画

%matplotlib inline

%config InlineBackend.figure_format = 'retina'

import pandas as pd

df = pd.read_csv(

filepath_or_buffer='https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data',

header=None,

sep=',')

df.columns=['sepal_len', 'sepal_wid', 'petal_len', 'petal_wid', 'class']

df.dropna(how="all", inplace=True) # drops the empty line at file-end

df.head()

X = df.ix[:,0:4].values

y = df.ix[:,4].values

现在上面数据处理后，x是一个150 * 4 的矩阵，每一行都是一个样本，y是一个 150 * 1 是向量，每个都是一个分类

我们下一步是来看3类型的花怎么分布在4个特征上，我们可以通过直方图来展示

import plotly.plotly as py

from plotly.graph_objs import *

import plotly.tools as tls

# plotting histograms

tls.set_credentials_file(username='zhuanxuhit', api_key='30dCVmghG2CqKQqfSzsu')

traces = []

legend = {0:False, 1:False, 2:False, 3:True}

colors = {'Iris-setosa': 'rgb(31, 119, 180)',

'Iris-versicolor': 'rgb(255, 127, 14)',

'Iris-virginica': 'rgb(44, 160, 44)'}

for col in range(4):

for key in colors:

traces.append(Histogram(x=X[y==key, col],

opacity=0.75,

xaxis='x%s' %(col+1),

marker=Marker(color=colors[key]),

name=key,

showlegend=legend[col]))

data = Data(traces)

layout = Layout(barmode='overlay',

xaxis=XAxis(domain=[0, 0.25], title='sepal length (cm)'),

xaxis2=XAxis(domain=[0.3, 0.5], title='sepal width (cm)'),

xaxis3=XAxis(domain=[0.55, 0.75], title='petal length (cm)'),

xaxis4=XAxis(domain=[0.8, 1], title='petal width (cm)'),

yaxis=YAxis(title='count'),

title='Distribution of the differ