专业建设网站的企业,做设计什么兼职网站建设,wordpress侧边浮动,长沙网站优化步骤目录 文章导航一、EDA#xff1a;二、导入类库三、导入数据四、查看数据类型和缺失情况五、确认目标变量和ID六、查看目标变量分布情况七、特征变量按照数据类型分成定量变量和定性变量八、查看定量变量分布情况九、查看定量变量的离散程度十、查看定量变量与目标变量关系十一… 目录 文章导航一、EDA二、导入类库三、导入数据四、查看数据类型和缺失情况五、确认目标变量和ID六、查看目标变量分布情况七、特征变量按照数据类型分成定量变量和定性变量八、查看定量变量分布情况九、查看定量变量的离散程度十、查看定量变量与目标变量关系十一、查看定性变量分布情况十二、查看定性变量与目标变量关系十三、查看定性变量对目标变量的显著性影响十四、查看定性变量和目标变量的spearman相关系数十五、查看定量变量与目标变量相关性十六、查看定性变量与目标变量相关性 文章导航
【一 简明数据分析进阶路径介绍文章导航】
一、EDA
EDAExploratory Data Analysis即探索性数据分析EDA通过可视化、统计和图形化的方法对数据集进行全面的、非形式化的初步分析帮助分析人员了解数据的基本特征发现数据中的规律和模式。这有助于获取对数据的直观感受和深刻理解为后续的数据处理和建模提供基础。
二、导入类库
# 导入类库
import numpy as np
import pandas as pd
import scipy.stats as statsimport matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px import warnings
warnings.filterwarnings(ignore)
from sklearn.preprocessing import LabelEncoder
from sklearn.preprocessing import RobustScalerfrom sklearn.decomposition import PCA
from sklearn.model_selection import cross_val_score, GridSearchCV, KFoldfrom sklearn.base import BaseEstimator, TransformerMixin, RegressorMixin
from sklearn.base import clone
from sklearn.linear_model import Lasso
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import Ridge
from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor, ExtraTreesRegressor
from sklearn.svm import SVR, LinearSVR
from sklearn.linear_model import ElasticNet, SGDRegressor, BayesianRidge
from sklearn.kernel_ridge import KernelRidge
from xgboost import XGBRegressor
# 显示中文
plt.rcParams[font.sans-serif] [SimHei]
plt.rcParams[axes.unicode_minus] False# pandas显示所有行和列
pd.set_option(display.max_columns, None)
pd.set_option(display.max_rows, None)三、导入数据
train pd.read_csv(./train.csv)
test pd.read_csv(./test.csv)train.head()四、查看数据类型和缺失情况
train.info()class pandas.core.frame.DataFrame
RangeIndex: 90615 entries, 0 to 90614
Data columns (total 10 columns):# Column Non-Null Count Dtype
--- ------ -------------- ----- 0 id 90615 non-null int64 1 Sex 90615 non-null object 2 Length 90615 non-null float643 Diameter 90615 non-null float644 Height 90615 non-null float645 Whole weight 90615 non-null float646 Whole weight.1 90615 non-null float647 Whole weight.2 90615 non-null float648 Shell weight 90615 non-null float649 Rings 90615 non-null int64
dtypes: float64(7), int64(2), object(1)
memory usage: 6.9 MB五、确认目标变量和ID
Target_features [Rings] #目标变量
ID_features [id] #id六、查看目标变量分布情况
Target_counts train[Target_features].value_counts().reset_index()
Target_counts.columns [Target_features[0], Count] # 绘制条形图
fig px.bar(Target_counts,xTarget_features[0], yCount, titleTarget_features[0]分布) # 遍历每个轨迹并设置文本
def set_text(trace): trace.text [f{val:.1f} for val in trace.y] trace.textposition outside fig.for_each_trace(set_text) # 显示图表
fig.show()七、特征变量按照数据类型分成定量变量和定性变量
# 移除ID和目标变量
train_columns list(train.columns)
train_columns.remove(Target_features[0])
train_columns.remove(ID_features[0])# 特征变量按照数据类型分成定量变量和定性变量
quantitative [feature for feature in train_columns if train.dtypes[feature] ! object] # 定量变量
print(定量变量)
print(quantitative)
qualitative [feature for feature in train_columns if train.dtypes[feature] object] # 定性变量
print(定性变量)
print(qualitative)定量变量
[Length, Diameter, Height, Whole weight, Whole weight.1, Whole weight.2, Shell weight]
定性变量
[Sex]八、查看定量变量分布情况
# 查看定量变量分布情况
m_cont pd.melt(train, value_varsquantitative)
g sns.FacetGrid(m_cont, colvariable, col_wrap4, sharexFalse, shareyFalse)
g.map(sns.distplot, value)九、查看定量变量的离散程度
# 查看定量变量的离散程度
def plot_boxplots(df):m_disc pd.melt(df)g sns.FacetGrid(m_disc, colvariable, col_wrap4, sharexFalse, shareyFalse)g.map(sns.boxplot, variable, value, width0.5)plt.show()plot_boxplots(train[quantitative]) 十、查看定量变量与目标变量关系
# 定量变量与目标变量关系图
m_cont pd.melt(train, id_varsTarget_features[0], value_varsquantitative)
g sns.FacetGrid(m_cont, colvariable, col_wrap4, sharexFalse, shareyTrue)
g.map(plt.scatter, value, Target_features[0])十一、查看定性变量分布情况
# 定性变量频数统计图
m_disc pd.melt(train, value_varsqualitative)
g sns.FacetGrid(m_disc, colvariable, col_wrap4, sharexFalse, shareyFalse)
g.map(sns.countplot, value)十二、查看定性变量与目标变量关系
# 定性变量与目标变量关系图
m_disc pd.melt(train, id_varsTarget_features[0], value_varsqualitative)
g sns.FacetGrid(m_disc, colvariable, col_wrap4, sharexFalse, shareyFalse)
g.map(sns.boxplot, value, Target_features[0])十三、查看定性变量对目标变量的显著性影响
# 查看定性变量对目标变量的显著性影响
def anova(frame, qualitative):anv pd.DataFrame()anv[feature] qualitativep_vals []for fea in qualitative:samples []cls frame[fea].unique() # 变量的类别值for c in cls:c_array frame[frame[fea]c][Target_features[0]].valuessamples.append(c_array)p_val stats.f_oneway(*samples)[1] # 获得p值,p值越小对SalePrice的显著性影响越大p_vals.append(p_val)anv[pval] p_valsreturn anv.sort_values(pval)
a anova(train, qualitative)
a[disparity] np.log(1./a[pval].values) # 对SalePrice的影响悬殊度
plt.figure(figsize(8, 6))
sns.barplot(xfeature, ydisparity, dataa)
plt.xticks(rotation90)
plt.show()十四、查看定性变量和目标变量的spearman相关系数
# 查看定性变量和目标变量的spearman相关系数
# 需要先把定性变量处理为数值类型
def encode(frame, feature):ordering pd.DataFrame()ordering[val] frame[feature].unique()ordering.index ordering[val]ordering[spmean] frame[[feature, Target_features[0]]].groupby(feature)[Target_features[0]].mean()ordering ordering.sort_values(spmean)ordering[ordering] np.arange(1, ordering.shape[0]1)ordering ordering[ordering].to_dict() # 返回的数据样例{category1:1, category2:2, ...}# 对frame[feature]编码for category, code_value in ordering.items():frame.loc[frame[feature]category, feature_E] code_value
qual_encoded []
for qual in qualitative:encode(train, qual)qual_encoded.append(qual_E)
# print(qual_encoded)def spearman(frame, features):spr pd.DataFrame()spr[feature] featuresspr[spearman] [frame[f].corr(frame[Target_features[0]], spearman) for f in features]spr spr.sort_values(spearman)plt.figure(figsize(6, 0.25*len(features)))sns.barplot(xspearman, yfeature, dataspr)
spearman(train, quantitativequal_encoded)十五、查看定量变量与目标变量相关性
# 定量变量与目标变量相关性
# plt.figure(1, figsize(12,9))
corrmat train[quantitative[Target_features[0]]].corr()
k 10 #number of variables for heatmap
cols corrmat.nlargest(k, Target_features[0])[Target_features[0]].index
corr train[list(cols)].corr()
sns.set(font_scale1.25)
sns.heatmap(corr, cbarTrue, annotTrue, squareTrue, fmt.2f, annot_kws{size: 10}, yticklabelscols.values, xticklabelscols.values)
plt.show()十六、查看定性变量与目标变量相关性
# 定性变量与目标变量相关性# plt.figure(1, figsize(12,9))
corrmat train[qual_encoded[Target_features[0]]].corr()
k 10 #number of variables for heatmap
cols corrmat.nlargest(k, Target_features[0])[Target_features[0]].index
corr train[list(cols)].corr()
sns.set(font_scale1.25)
sns.heatmap(corr, cbarTrue, annotTrue, squareTrue, fmt.2f, annot_kws{size: 10}, yticklabelscols.values, xticklabelscols.values)
plt.show()
